Search results for: implicit neural representations
Commenced in January 2007
Frequency: Monthly
Edition: International
Paper Count: 2336

Search results for: implicit neural representations

296 The Development of Modernist Chinese Architecture from the Perspective of Cultural Regionalism in Taiwan: Spatial Practice by the Fieldoffice Architects

Authors: Yilei Yu

Abstract:

Modernism, emerging in the Western world of the 20th century, attempted to create a universal international style, which pulled the architectural and social systems created by classicism back to an initial pure state. However, out of the introspection of the Modernism, Regionalism attempted to restore a humanistic environment and create flexible buildings during the 1950s. Meanwhile, as the first generation of architects came back, the wind of the Regionalism blew to Taiwan. However, with the increasing of political influence and the tightening of free creative space, from the second half of the 1950s to the 1980s, the "real" Regional Architecture, which should have taken roots in Taiwan, becomes the "fake" Regional Architecture filled with the oriental charm. Through the Comparative Method, which includes description, interpretation, juxtaposition, and comparison, this study analyses the difference of the style of the Modernist Chinese Architecture between the period before the 1980s and the after. The paper aims at exploring the development of Regionalism Architecture in Taiwan, which includes three parts. First, the burgeoning period of the "modernist Chinese architecture" in Taiwan was the beginning of the Chinese Nationalist Party's coming to Taiwan to consolidate political power. The architecture of the "Ming and Qing Dynasty Palace Revival Style" dominated the architectural circles in Taiwan. These superficial "regional buildings" have nearly no combination with the local customs of Taiwan, which is difficult to evoke the social identity. Second, in the late 1970s, the second generation of architects headed by Baode Han began focusing on the research and preservation of traditional Taiwanese architecture, and creating buildings combined the terroirs of Taiwan through the imitation of styles. However, some scholars have expressed regret that very few regionalist architectural works that appeared in the 1980s can respond specifically to regional conditions and forms of construction. Instead, most of them are vocabulary-led representations. Third, during the 1990s, by the end of the period of martial law, community building gradually emerged, which made the object of Taiwan's architectural concern gradually extended to the folk and ethnic groups. In the Yilan area, there are many architects who care about the local environment, such as the Field office Architects. Compared with the hollow regionality of the passionate national spirits that emerged during the martial law period, the local practice of the architect team in Yilan can better link the real local environmental life and reflect the true regionality. In conclusion, with the local practice case of the huge construction team in Yilan area, this paper focuses on the Spatial Practice by the Field office Architects to explore the spatial representation of the space and the practical enlightenment in the process of modernist Chinese architecture development in Taiwan.

Keywords: regionalism, modernism, Chinese architecture, political landscape, spatial representation

Procedia PDF Downloads 106
295 Multimodal Sentiment Analysis With Web Based Application

Authors: Shreyansh Singh, Afroz Ahmed

Abstract:

Sentiment Analysis intends to naturally reveal the hidden mentality that we hold towards an entity. The total of this assumption over a populace addresses sentiment surveying and has various applications. Current text-based sentiment analysis depends on the development of word embeddings and Machine Learning models that take in conclusion from enormous text corpora. Sentiment Analysis from text is presently generally utilized for consumer loyalty appraisal and brand insight investigation. With the expansion of online media, multimodal assessment investigation is set to carry new freedoms with the appearance of integral information streams for improving and going past text-based feeling examination using the new transforms methods. Since supposition can be distinguished through compelling follows it leaves, like facial and vocal presentations, multimodal opinion investigation offers good roads for examining facial and vocal articulations notwithstanding the record or printed content. These methodologies use the Recurrent Neural Networks (RNNs) with the LSTM modes to increase their performance. In this study, we characterize feeling and the issue of multimodal assessment investigation and audit ongoing advancements in multimodal notion examination in various spaces, including spoken surveys, pictures, video websites, human-machine, and human-human connections. Difficulties and chances of this arising field are additionally examined, promoting our theory that multimodal feeling investigation holds critical undiscovered potential.

Keywords: sentiment analysis, RNN, LSTM, word embeddings

Procedia PDF Downloads 96
294 Fake News Detection for Korean News Using Machine Learning Techniques

Authors: Tae-Uk Yun, Pullip Chung, Kee-Young Kwahk, Hyunchul Ahn

Abstract:

Fake news is defined as the news articles that are intentionally and verifiably false, and could mislead readers. Spread of fake news may provoke anxiety, chaos, fear, or irrational decisions of the public. Thus, detecting fake news and preventing its spread has become very important issue in our society. However, due to the huge amount of fake news produced every day, it is almost impossible to identify it by a human. Under this context, researchers have tried to develop automated fake news detection using machine learning techniques over the past years. But, there have been no prior studies proposed an automated fake news detection method for Korean news to our best knowledge. In this study, we aim to detect Korean fake news using text mining and machine learning techniques. Our proposed method consists of two steps. In the first step, the news contents to be analyzed is convert to quantified values using various text mining techniques (topic modeling, TF-IDF, and so on). After that, in step 2, classifiers are trained using the values produced in step 1. As the classifiers, machine learning techniques such as logistic regression, backpropagation network, support vector machine, and deep neural network can be applied. To validate the effectiveness of the proposed method, we collected about 200 short Korean news from Seoul National University’s FactCheck. which provides with detailed analysis reports from 20 media outlets and links to source documents for each case. Using this dataset, we will identify which text features are important as well as which classifiers are effective in detecting Korean fake news.

Keywords: fake news detection, Korean news, machine learning, text mining

Procedia PDF Downloads 247
293 The Curvature of Bending Analysis and Motion of Soft Robotic Fingers by Full 3D Printing with MC-Cells Technique for Hand Rehabilitation

Authors: Chaiyawat Musikapan, Ratchatin Chancharoen, Saknan Bongsebandhu-Phubhakdi

Abstract:

For many recent years, soft robotic fingers were used for supporting the patients who had survived the neurological diseases that resulted in muscular disorders and neural network damages, such as stroke and Parkinson’s disease, and inflammatory symptoms such as De Quervain and trigger finger. Generally, the major hand function is significant to manipulate objects in activities of daily living (ADL). In this work, we proposed the model of soft actuator that manufactured by full 3D printing without the molding process and one material for use. Furthermore, we designed the model with a technique of multi cavitation cells (MC-Cells). Then, we demonstrated the curvature bending, fluidic pressure and force that generated to the model for assistive finger flexor and hand grasping. Also, the soft actuators were characterized in mathematics solving by the length of chord and arc length. In addition, we used an adaptive push-button switch machine to measure the force in our experiment. Consequently, we evaluated biomechanics efficiency by the range of motion (ROM) that affected to metacarpophalangeal joint (MCP), proximal interphalangeal joint (PIP) and distal interphalangeal joint (DIP). Finally, the model achieved to exhibit the corresponding fluidic pressure with force and ROM to assist the finger flexor and hand grasping.

Keywords: biomechanics efficiency, curvature bending, hand functional assistance, multi cavitation cells (MC-Cells), range of motion (ROM)

Procedia PDF Downloads 232
292 Refined Edge Detection Network

Authors: Omar Elharrouss, Youssef Hmamouche, Assia Kamal Idrissi, Btissam El Khamlichi, Amal El Fallah-Seghrouchni

Abstract:

Edge detection is represented as one of the most challenging tasks in computer vision, due to the complexity of detecting the edges or boundaries in real-world images that contains objects of different types and scales like trees, building as well as various backgrounds. Edge detection is represented also as a key task for many computer vision applications. Using a set of backbones as well as attention modules, deep-learning-based methods improved the detection of edges compared with the traditional methods like Sobel and Canny. However, images of complex scenes still represent a challenge for these methods. Also, the detected edges using the existing approaches suffer from non-refined results while the image output contains many erroneous edges. To overcome this, n this paper, by using the mechanism of residual learning, a refined edge detection network is proposed (RED-Net). By maintaining the high resolution of edges during the training process, and conserving the resolution of the edge image during the network stage, we make the pooling outputs at each stage connected with the output of the previous layer. Also, after each layer, we use an affined batch normalization layer as an erosion operation for the homogeneous region in the image. The proposed methods are evaluated using the most challenging datasets including BSDS500, NYUD, and Multicue. The obtained results outperform the designed edge detection networks in terms of performance metrics and quality of output images.

Keywords: edge detection, convolutional neural networks, deep learning, scale-representation, backbone

Procedia PDF Downloads 78
291 Ant Lion Optimization in a Fuzzy System for Benchmark Control Problem

Authors: Leticia Cervantes, Edith Garcia, Oscar Castillo

Abstract:

At today, there are several control problems where the main objective is to obtain the best control in the study to decrease the error in the application. Many techniques can use to control these problems such as Neural Networks, PID control, Fuzzy Logic, Optimization techniques and many more. In this case, fuzzy logic with fuzzy system and an optimization technique are used to control the case of study. In this case, Ant Lion Optimization is used to optimize a fuzzy system to control the velocity of a simple treadmill. The main objective is to achieve the control of the velocity in the control problem using the ALO optimization. First, a simple fuzzy system was used to control the velocity of the treadmill it has two inputs (error and error change) and one output (desired speed), then results were obtained but to decrease the error the ALO optimization was developed to optimize the fuzzy system of the treadmill. Having the optimization, the simulation was performed, and results can prove that using the ALO optimization the control of the velocity was better than a conventional fuzzy system. This paper describes some basic concepts to help to understand the idea in this work, the methodology of the investigation (control problem, fuzzy system design, optimization), the results are presented and the optimization is used for the fuzzy system. A comparison between the simple fuzzy system and the optimized fuzzy systems are presented where it can be proving the optimization improved the control with good results the major findings of the study is that ALO optimization is a good alternative to improve the control because it helped to decrease the error in control applications even using any control technique to optimized, As a final statement is important to mentioned that the selected methodology was good because the control of the treadmill was improve using the optimization technique.

Keywords: ant lion optimization, control problem, fuzzy control, fuzzy system

Procedia PDF Downloads 370
290 Supporting a Moral Growth Mindset Among College Students

Authors: Kate Allman, Heather Maranges, Elise Dykhuis

Abstract:

Moral Growth Mindset (MGM) is the belief that one has the capacity to become a more moral person, as opposed to a fixed conception of one’s moral ability and capacity (Han et al., 2018). Building from Dweck’s work in incremental implicit theories of intelligence (2008), Moral Growth Mindset (Han et al., 2020) extends growth mindsets into the moral dimension. The concept of MGM has the potential to help researchers understand how both mindsets and interventions can impact character development, and it has even been shown to have connections to voluntary service engagement (Han et al., 2018). Understanding the contexts in which MGM might be cultivated could help to promote the further cultivation of character, in addition to prosocial behaviors like service engagement, which may, in turn, promote larger scale engagement in social justice-oriented thoughts, feelings, and behaviors. In particular, college may be a place to intentionally cultivate a growth mindset toward moral capacities, given the unique developmental and maturational components of the college experience, including contextual opportunity (Lapsley & Narvaez, 2006) and independence requiring the constant consideration, revision, and internalization of personal values (Lapsley & Woodbury, 2016). In a semester-long, quasi-experimental study, we examined the impact of a pedagogical approach designed to cultivate college student character development on participants’ MGM. With an intervention (n=69) and a control group (n=97; Pre-course: 27% Men; 66% Women; 68% White; 18% Asian; 2% Black; <1% Hispanic/Latino), we investigated whether college courses that intentionally incorporate character education pedagogy (Lamb, Brant, Brooks, 2021) affect a variety of psychosocial variables associated with moral thoughts, feelings, identity, and behavior (e.g. moral growth mindset, honesty, compassion, etc.). The intervention group consisted of 69 undergraduate students (Pre-course: 40% Men; 52% Women; 68% White; 10.5% Black; 7.4% Asian; 4.2% Hispanic/Latino) that voluntarily enrolled in five undergraduate courses that encouraged students to engage with key concepts and methods of character development through the application of research-based strategies and personal reflection on goals and experiences. Moral Growth Mindset was measured using the four-item Moral Growth Mindset scale (Han et al., 2020), with items such as You can improve your basic morals and character considerably on a six-point Likert scale from 1 (strongly disagree) to 6 (strongly agree). Higher scores of MGM indicate a stronger belief that one can become a more moral person with personal effort. Reliability at Time 1 was Cronbach’s ɑ= .833, and at Time 2 Cronbach’s ɑ= .772. An Analysis of Covariance (ANCOVA) was conducted to explore whether post-course MGM scores were different between the intervention and control when controlling for pre-course MGM scores. The ANCOVA indicated significant differences in MGM between groups post-course, F(1,163) = 8.073, p = .005, R² = .11, where descriptive statistics indicate that intervention scores were higher than the control group at post-course. Results indicate that intentional character development pedagogy can be leveraged to support the development of Moral Growth Mindset and related capacities in undergraduate settings.

Keywords: moral personality, character education, incremental theories of personality, growth mindset

Procedia PDF Downloads 126
289 Advancing Our Understanding of Age-Related Changes in Executive Functions: Insights from Neuroimaging, Genetics and Cognitive Neurosciences

Authors: Yasaman Mohammadi

Abstract:

Executive functions are a critical component of goal-directed behavior, encompassing a diverse set of cognitive processes such as working memory, cognitive flexibility, and inhibitory control. These functions are known to decline with age, but the precise mechanisms underlying this decline remain unclear. This paper provides an in-depth review of recent research investigating age-related changes in executive functions, drawing on insights from neuroimaging, genetics, and cognitive neuroscience. Through an interdisciplinary approach, this paper offers a nuanced understanding of the complex interplay between neural mechanisms, genetic factors, and cognitive processes that contribute to executive function decline in aging. Here, we investigate how different neuroimaging methods, like functional magnetic resonance imaging (fMRI) and positron emission tomography (PET), have helped scientists better understand the brain bases for age-related declines in executive function. Additionally, we discuss the role of genetic factors in mediating individual differences in executive functions across the lifespan, as well as the potential for cognitive interventions to mitigate age-related decline. Overall, this paper presents a comprehensive and integrative view of the current state of knowledge regarding age-related changes in executive functions. It underscores the need for continued interdisciplinary research to fully understand the complex and dynamic nature of executive function decline in aging, with the ultimate goal of developing effective interventions to promote healthy cognitive aging.

Keywords: executive functions, aging, neuroimaging, cognitive neuroscience, working memory, cognitive training

Procedia PDF Downloads 47
288 A Lexicographic Approach to Obstacles Identified in the Ontological Representation of the Tree of Life

Authors: Sandra Young

Abstract:

The biodiversity literature is vast and heterogeneous. In today’s data age, numbers of data integration and standardisation initiatives aim to facilitate simultaneous access to all the literature across biodiversity domains for research and forecasting purposes. Ontologies are being used increasingly to organise this information, but the rationalisation intrinsic to ontologies can hit obstacles when faced with the intrinsic fluidity and inconsistency found in the domains comprising biodiversity. Essentially the problem is a conceptual one: biological taxonomies are formed on the basis of specific, physical specimens yet nomenclatural rules are used to provide labels to describe these physical objects. These labels are ambiguous representations of the physical specimen. An example of this is with the genus Melpomene, the scientific nomenclatural representation of a genus of ferns, but also for a genus of spiders. The physical specimens for each of these are vastly different, but they have been assigned the same nomenclatural reference. While there is much research into the conceptual stability of the taxonomic concept versus the nomenclature used, to the best of our knowledge as yet no research has looked empirically at the literature to see the conceptual plurality or singularity of the use of these species’ names, the linguistic representation of a physical entity. Language itself uses words as symbols to represent real world concepts, whether physical entities or otherwise, and as such lexicography has a well-founded history in the conceptual mapping of words in context for dictionary making. This makes it an ideal candidate to explore this problem. The lexicographic approach uses corpus-based analysis to look at word use in context, with a specific focus on collocated word frequencies (the frequencies of words used in specific grammatical and collocational contexts). It allows for inconsistencies and contradictions in the source data and in fact includes these in the word characterisation so that 100% of the available evidence is counted. Corpus analysis is indeed suggested as one of the ways to identify concepts for ontology building, because of its ability to look empirically at data and show patterns in language usage, which can indicate conceptual ideas which go beyond words themselves. In this sense it could potentially be used to identify if the hierarchical structures present within the empirical body of literature match those which have been identified in ontologies created to represent them. The first stages of this research have revealed a hierarchical structure that becomes apparent in the biodiversity literature when annotating scientific species’ names, common names and more general names as classes, which will be the focus of this paper. The next step in the research is focusing on a larger corpus in which specific words can be analysed and then compared with existing ontological structures looking at the same material, to evaluate the methods by means of an alternative perspective. This research aims to provide evidence as to the validity of the current methods in knowledge representation for biological entities, and also shed light on the way that scientific nomenclature is used within the literature.

Keywords: ontology, biodiversity, lexicography, knowledge representation, corpus linguistics

Procedia PDF Downloads 116
287 Exploring the Use of Augmented Reality for Laboratory Lectures in Distance Learning

Authors: Michele Gattullo, Vito M. Manghisi, Alessandro Evangelista, Enricoandrea Laviola

Abstract:

In this work, we explored the use of Augmented Reality (AR) to support students in laboratory lectures in Distance Learning (DL), designing an application that proved to be ready for use next semester. AR could help students in the understanding of complex concepts as well as increase their motivation in the learning process. However, despite many prototypes in the literature, it is still less used in schools and universities. This is mainly due to the perceived limited advantages to the investment costs, especially regarding changes needed in the teaching modalities. However, with the spread of epidemiological emergency due to SARS-CoV-2, schools and universities were forced to a very rapid redefinition of consolidated processes towards forms of Distance Learning. Despite its many advantages, it suffers from the impossibility to carry out practical activities that are of crucial importance in STEM ("Science, Technology, Engineering e Math") didactics. In this context, AR perceived advantages increased a lot since teachers are more prepared for new teaching modalities, exploiting AR that allows students to carry on practical activities on their own instead of being physically present in laboratories. In this work, we designed an AR application for the support of engineering students in the understanding of assembly drawings of complex machines. Traditionally, this skill is acquired in the first years of the bachelor's degree in industrial engineering, through laboratory activities where the teacher shows the corresponding components (e.g., bearings, screws, shafts) in a real machine and their representation in the assembly drawing. This research aims to explore the effectiveness of AR to allow students to acquire this skill on their own without physically being in the laboratory. In a preliminary phase, we interviewed students to understand the main issues in the learning of this subject. This survey revealed that students had difficulty identifying machine components in an assembly drawing, matching between the 2D representation of a component and its real shape, and understanding the functionality of a component within the machine. We developed a mobile application using Unity3D, aiming to solve the mentioned issues. We designed the application in collaboration with the course professors. Natural feature tracking was used to associate the 2D printed assembly drawing with the corresponding 3D virtual model. The application can be displayed on students’ tablets or smartphones. Users could interact with selecting a component from a part list on the device. Then, 3D representations of components appear on the printed drawing, coupled with 3D virtual labels for their location and identification. Users could also interact with watching a 3D animation to learn how components are assembled. Students evaluated the application through a questionnaire based on the System Usability Scale (SUS). The survey was provided to 15 students selected among those we participated in the preliminary interview. The mean SUS score was 83 (SD 12.9) over a maximum of 100, allowing teachers to use the AR application in their courses. Another important finding is that almost all the students revealed that this application would provide significant power for comprehension on their own.

Keywords: augmented reality, distance learning, STEM didactics, technology in education

Procedia PDF Downloads 108
286 Contraceptives: Experiences of Agency and Coercion of Young People Living in Colombia

Authors: Paola Montenegro, Maria de los Angeles Balaguera Villa

Abstract:

Contraceptive methods play a fundamental role in preventing unwanted pregnancies and protecting users from sexually transmitted infections (STIs). Despite being known to almost the entire population of reproductive age living in Colombia, there are barriers, practices and complex notions about contraceptives that affect their desired mass use and effectiveness. This work aims to analyse some of the perceptions and practices discussed with young people (13-28 years old) living in Colombia regarding the use of contraceptives in their daily lives, preferences, needs and perceived side effects. This research also examines the perceived paradox in autonomy that young people experience regarding contraceptive use: in one hand, its use (or lack of it) is interpreted as an act of self-determination and primary example of reproductive agency, on the other hand, it was frequently associated with coercion and limited autonomy derived from the gaps in reliable information available for young people, the difficulty of accessing certain preferred methods, and sometimes the experienced coercion exercise by doctors, partners and/or family members. The data and analysis discussed in this work stems from a research project whose objective was to provide information about needs and preferences in sexual and reproductive health of young people living in Colombia in relation to a possible telehealth service that could close the gap in access to quality care and safe information. Through a mixed methods approach, this study collected 5.736 responses to a virtual survey disseminated nationwide in Colombia and 47 inperson interviews (24 of them with people who were assigned female at birth and 21 with local key stakeholders in the abortion ecosystem). Quantitative data was analyzed using Stata SE Version 16.0 and qualitative analysis was completed through NVivo using thematic analysis. Key findings on contraception use in young people living in Colombia reveal that 85,8% of participants had used a contraceptive method in the last two years, and that the most commonly used methods were condoms, contraceptive pills, the morning-after pill and the method of interruption. The remaining 14,2% of respondents who declared to not have used contraceptives in the last two years expressed that the main four barriers to access were: "Lack of knowledge about contraceptive methods and where to obtain information and/or access them (13.9%)", "Have had sex with people who have vaginas (10.2%)", "Cost of contraceptive method (8.4%)" and "Difficulties in obtaining medical authorisations (7.6%)". These barriers coincided with the ones used to explain the non-use of contraceptives in young people, which reveals that limitations in information, cost, and quality care represent structural issues that need to be address in programmes, services, and public policy. Finally, interviews showed that young people perceive contraceptive use and non-use as an example of reaffirming reproductive agency and limitations to this can be explained through the widespread incomplete knowledge about how methods work and the prevalence of other social representations of contraception associated with trust, fidelity, and partner preferences, that in the end create limitations to young people’s autonomy.

Keywords: contraception, family planning, premarital fertility, unplanned pregnancy

Procedia PDF Downloads 50
285 The Readaptation of the Subscale 3 of the NLit-IT (Nutrition Literacy Assessment Instrument for Italian Subjects)

Authors: Virginia Vettori, Chiara Lorini, Vieri Lastrucci, Giulia Di Pisa, Alessia De Blasi, Sara Giuggioli, Guglielmo Bonaccorsi

Abstract:

The design of the Nutrition Literacy Assessment Instrument (NLit) responds to the need to provide a tool to adequately assess the construct of nutrition literacy (NL), which is strictly connected to the quality of the diet and nutritional health status. The NLit was originally developed and validated in the US context, and it was recently validated for Italian people too (NLit-IT), involving a sample of N = 74 adults. The results of the cross-cultural adaptation of the tool confirmed its validity since it was established that the level of NL contributed to predicting the level of adherence to the Mediterranean Diet (convergent validity). Additionally, results obtained proved that Internal Consistency and reliability of the NLit-IT were good (Cronbach’s alpha (ρT) = 0.78; 95% CI, 0.69–0.84; Intraclass Correlation Coefficient (ICC) = 0.68, 95% CI, 0.46–0.85). However, the Subscale 3 of the NLit-IT “Household Food Measurement” showed lower values of ρT and ICC (ρT = 0.27; 95% CI, 0.1–0.55; ICC = 0.19, 95% CI, 0.01–0.63) than the entire instrument. Subscale 3 includes nine items which are constituted by written questions and the corresponding pictures of the meals. In particular, items 2, 3, and 8 of Subscale 3 had the lowest level of correct answers. The purpose of the present study was to identify the factors that influenced the Internal Consistency and reliability of Subscale 3 of NLit-IT using the methodology of a focus group. A panel of seven experts was formed, involving professionals in the field of public health nutrition, dietetics, and health promotion and all of them were trained on the concepts of nutrition literacy and food appearance. A member of the group drove the discussion, which was oriented in the identification of the reasons for the low levels of reliability and Internal Consistency. The members of the group discussed the level of comprehension of the items and how they could be readapted. From the discussion, it emerges that the written questions were clear and easy to understand, but it was observed that the representations of the meal needed to be improved. Firstly, it has been decided to introduce a fork or a spoon as a reference dimension to better understand the dimension of the food portion (items 1, 4 and 8). Additionally, the flat plate of items 3 and 5 should be substituted with a soup plate because, in the Italian national context, it is common to eat pasta or rice on this kind of plate. Secondly, specific measures should be considered for some kind of foods such as the brick of yogurt instead of a cup of yogurt (items 1 and 4). Lastly, it has been decided to redo the photos of the meals basing on professional photographic techniques. In conclusion, we noted that the graphical representation of the items strictly influenced the level of participants’ comprehension of the questions; moreover, the research group agreed that the level of knowledge about nutrition and food portion size is low in the general population.

Keywords: nutritional literacy, cross cultural adaptation, misinformation, food design

Procedia PDF Downloads 142
284 Iterative Segmentation and Application of Hausdorff Dilation Distance in Defect Detection

Authors: S. Shankar Bharathi

Abstract:

Inspection of surface defects on metallic components has always been challenging due to its specular property. Occurrences of defects such as scratches, rust, pitting are very common in metallic surfaces during the manufacturing process. These defects if unchecked can hamper the performance and reduce the life time of such component. Many of the conventional image processing algorithms in detecting the surface defects generally involve segmentation techniques, based on thresholding, edge detection, watershed segmentation and textural segmentation. They later employ other suitable algorithms based on morphology, region growing, shape analysis, neural networks for classification purpose. In this paper the work has been focused only towards detecting scratches. Global and other thresholding techniques were used to extract the defects, but it proved to be inaccurate in extracting the defects alone. However, this paper does not focus on comparison of different segmentation techniques, but rather describes a novel approach towards segmentation combined with hausdorff dilation distance. The proposed algorithm is based on the distribution of the intensity levels, that is, whether a certain gray level is concentrated or evenly distributed. The algorithm is based on extraction of such concentrated pixels. Defective images showed higher level of concentration of some gray level, whereas in non-defective image, there seemed to be no concentration, but were evenly distributed. This formed the basis in detecting the defects in the proposed algorithm. Hausdorff dilation distance based on mathematical morphology was used to strengthen the segmentation of the defects.

Keywords: metallic surface, scratches, segmentation, hausdorff dilation distance, machine vision

Procedia PDF Downloads 397
283 An Integrated Framework for Seismic Risk Mitigation Decision Making

Authors: Mojtaba Sadeghi, Farshid Baniassadi, Hamed Kashani

Abstract:

One of the challenging issues faced by seismic retrofitting consultants and employers is quick decision-making on the demolition or retrofitting of a structure at the current time or in the future. For this reason, the existing models proposed by researchers have only covered one of the aspects of cost, execution method, and structural vulnerability. Given the effect of each factor on the final decision, it is crucial to devise a new comprehensive model capable of simultaneously covering all the factors. This study attempted to provide an integrated framework that can be utilized to select the most appropriate earthquake risk mitigation solution for buildings. This framework can overcome the limitations of current models by taking into account several factors such as cost, execution method, risk-taking and structural failure. In the newly proposed model, the database and essential information about retrofitting projects are developed based on the historical data on a retrofit project. In the next phase, an analysis is conducted in order to assess the vulnerability of the building under study. Then, artificial neural networks technique is employed to calculate the cost of retrofitting. While calculating the current price of the structure, an economic analysis is conducted to compare demolition versus retrofitting costs. At the next stage, the optimal method is identified. Finally, the implementation of the framework was demonstrated by collecting data concerning 155 previous projects.

Keywords: decision making, demolition, construction management, seismic retrofit

Procedia PDF Downloads 218
282 Enhancer: An Effective Transformer Architecture for Single Image Super Resolution

Authors: Pitigalage Chamath Chandira Peiris

Abstract:

A widely researched domain in the field of image processing in recent times has been single image super-resolution, which tries to restore a high-resolution image from a single low-resolution image. Many more single image super-resolution efforts have been completed utilizing equally traditional and deep learning methodologies, as well as a variety of other methodologies. Deep learning-based super-resolution methods, in particular, have received significant interest. As of now, the most advanced image restoration approaches are based on convolutional neural networks; nevertheless, only a few efforts have been performed using Transformers, which have demonstrated excellent performance on high-level vision tasks. The effectiveness of CNN-based algorithms in image super-resolution has been impressive. However, these methods cannot completely capture the non-local features of the data. Enhancer is a simple yet powerful Transformer-based approach for enhancing the resolution of images. A method for single image super-resolution was developed in this study, which utilized an efficient and effective transformer design. This proposed architecture makes use of a locally enhanced window transformer block to alleviate the enormous computational load associated with non-overlapping window-based self-attention. Additionally, it incorporates depth-wise convolution in the feed-forward network to enhance its ability to capture local context. This study is assessed by comparing the results obtained for popular datasets to those obtained by other techniques in the domain.

Keywords: single image super resolution, computer vision, vision transformers, image restoration

Procedia PDF Downloads 82
281 In and Out-Of-Sample Performance of Non Simmetric Models in International Price Differential Forecasting in a Commodity Country Framework

Authors: Nicola Rubino

Abstract:

This paper presents an analysis of a group of commodity exporting countries' nominal exchange rate movements in relationship to the US dollar. Using a series of Unrestricted Self-exciting Threshold Autoregressive models (SETAR), we model and evaluate sixteen national CPI price differentials relative to the US dollar CPI. Out-of-sample forecast accuracy is evaluated through calculation of mean absolute error measures on the basis of two-hundred and fifty-three months rolling window forecasts and extended to three additional models, namely a logistic smooth transition regression (LSTAR), an additive non linear autoregressive model (AAR) and a simple linear Neural Network model (NNET). Our preliminary results confirm presence of some form of TAR non linearity in the majority of the countries analyzed, with a relatively higher goodness of fit, with respect to the linear AR(1) benchmark, in five countries out of sixteen considered. Although no model appears to statistically prevail over the other, our final out-of-sample forecast exercise shows that SETAR models tend to have quite poor relative forecasting performance, especially when compared to alternative non-linear specifications. Finally, by analyzing the implied half-lives of the > coefficients, our results confirms the presence, in the spirit of arbitrage band adjustment, of band convergence with an inner unit root behaviour in five of the sixteen countries analyzed.

Keywords: transition regression model, real exchange rate, nonlinearities, price differentials, PPP, commodity points

Procedia PDF Downloads 257
280 A Picture is worth a Billion Bits: Real-Time Image Reconstruction from Dense Binary Pixels

Authors: Tal Remez, Or Litany, Alex Bronstein

Abstract:

The pursuit of smaller pixel sizes at ever increasing resolution in digital image sensors is mainly driven by the stringent price and form-factor requirements of sensors and optics in the cellular phone market. Recently, Eric Fossum proposed a novel concept of an image sensor with dense sub-diffraction limit one-bit pixels (jots), which can be considered a digital emulation of silver halide photographic film. This idea has been recently embodied as the EPFL Gigavision camera. A major bottleneck in the design of such sensors is the image reconstruction process, producing a continuous high dynamic range image from oversampled binary measurements. The extreme quantization of the Poisson statistics is incompatible with the assumptions of most standard image processing and enhancement frameworks. The recently proposed maximum-likelihood (ML) approach addresses this difficulty, but suffers from image artifacts and has impractically high computational complexity. In this work, we study a variant of a sensor with binary threshold pixels and propose a reconstruction algorithm combining an ML data fitting term with a sparse synthesis prior. We also show an efficient hardware-friendly real-time approximation of this inverse operator. Promising results are shown on synthetic data as well as on HDR data emulated using multiple exposures of a regular CMOS sensor.

Keywords: binary pixels, maximum likelihood, neural networks, sparse coding

Procedia PDF Downloads 179
279 Text Localization in Fixed-Layout Documents Using Convolutional Networks in a Coarse-to-Fine Manner

Authors: Beier Zhu, Rui Zhang, Qi Song

Abstract:

Text contained within fixed-layout documents can be of great semantic value and so requires a high localization accuracy, such as ID cards, invoices, cheques, and passports. Recently, algorithms based on deep convolutional networks achieve high performance on text detection tasks. However, for text localization in fixed-layout documents, such algorithms detect word bounding boxes individually, which ignores the layout information. This paper presents a novel architecture built on convolutional neural networks (CNNs). A global text localization network and a regional bounding-box regression network are introduced to tackle the problem in a coarse-to-fine manner. The text localization network simultaneously locates word bounding points, which takes the layout information into account. The bounding-box regression network inputs the features pooled from arbitrarily sized RoIs and refine the localizations. These two networks share their convolutional features and are trained jointly. A typical type of fixed-layout documents: ID cards, is selected to evaluate the effectiveness of the proposed system. These networks are trained on data cropped from nature scene images, and synthetic data produced by a synthetic text generation engine. Experiments show that our approach locates high accuracy word bounding boxes and achieves state-of-the-art performance.

Keywords: bounding box regression, convolutional networks, fixed-layout documents, text localization

Procedia PDF Downloads 170
278 Big Data in Telecom Industry: Effective Predictive Techniques on Call Detail Records

Authors: Sara ElElimy, Samir Moustafa

Abstract:

Mobile network operators start to face many challenges in the digital era, especially with high demands from customers. Since mobile network operators are considered a source of big data, traditional techniques are not effective with new era of big data, Internet of things (IoT) and 5G; as a result, handling effectively different big datasets becomes a vital task for operators with the continuous growth of data and moving from long term evolution (LTE) to 5G. So, there is an urgent need for effective Big data analytics to predict future demands, traffic, and network performance to full fill the requirements of the fifth generation of mobile network technology. In this paper, we introduce data science techniques using machine learning and deep learning algorithms: the autoregressive integrated moving average (ARIMA), Bayesian-based curve fitting, and recurrent neural network (RNN) are employed for a data-driven application to mobile network operators. The main framework included in models are identification parameters of each model, estimation, prediction, and final data-driven application of this prediction from business and network performance applications. These models are applied to Telecom Italia Big Data challenge call detail records (CDRs) datasets. The performance of these models is found out using a specific well-known evaluation criteria shows that ARIMA (machine learning-based model) is more accurate as a predictive model in such a dataset than the RNN (deep learning model).

Keywords: big data analytics, machine learning, CDRs, 5G

Procedia PDF Downloads 117
277 Effective Stacking of Deep Neural Models for Automated Object Recognition in Retail Stores

Authors: Ankit Sinha, Soham Banerjee, Pratik Chattopadhyay

Abstract:

Automated product recognition in retail stores is an important real-world application in the domain of Computer Vision and Pattern Recognition. In this paper, we consider the problem of automatically identifying the classes of the products placed on racks in retail stores from an image of the rack and information about the query/product images. We improve upon the existing approaches in terms of effectiveness and memory requirement by developing a two-stage object detection and recognition pipeline comprising of a Faster-RCNN-based object localizer that detects the object regions in the rack image and a ResNet-18-based image encoder that classifies the detected regions into the appropriate classes. Each of the models is fine-tuned using appropriate data sets for better prediction and data augmentation is performed on each query image to prepare an extensive gallery set for fine-tuning the ResNet-18-based product recognition model. This encoder is trained using a triplet loss function following the strategy of online-hard-negative-mining for improved prediction. The proposed models are lightweight and can be connected in an end-to-end manner during deployment to automatically identify each product object placed in a rack image. Extensive experiments using Grozi-32k and GP-180 data sets verify the effectiveness of the proposed model.

Keywords: retail stores, faster-RCNN, object localization, ResNet-18, triplet loss, data augmentation, product recognition

Procedia PDF Downloads 124
276 Comparative Study of Properties of Iranian Historical Gardens by Focusing on Climate

Authors: Malihe Ahmadi

Abstract:

Nowadays, stress, tension and neural problems are among the most important concerns of the present age. The environment plays key role on improving mental health and reducing stress of citizens. Establishing balance and appropriate relationship between city and natural environment is of the most important approaches of present century. Type of approach and logical planning for urban green spaces as one of the basic sections of integration with nature, not only plays key role on quality and efficiency of comprehensive urban planning; but also it increases the system of distributing social activities and happiness and lively property of urban environments that leads to permanent urban development. The main purpose of recovering urban identity is considering culture, history and human life style in past. This is a documentary-library research that evaluates the historical properties of Iranian gardens in compliance with climate condition. Results of this research reveal that in addition to following Iranian gardens from common principles of land lot, structure of flowers and plants, water, specific buildings during different ages, the role of climate at different urban areas is among the basics of determining method of designing green spaces and different buildings located at diverse areas i.e. Iranian gardens are a space for merging natural and artificial elements that has inseparable connection with semantic principles and guarantees different functions. Some of the necessities of designing present urban gardens are including: recognition and recreation.

Keywords: historical gardens, climate, properties of Iranian gardens, Iran

Procedia PDF Downloads 378
275 The Home as Memory Palace: Three Case Studies of Artistic Representations of the Relationship between Individual and Collective Memory and the Home

Authors: Laura M. F. Bertens

Abstract:

The houses we inhabit are important containers of memory. As homes, they take on meaning for those who live inside, and memories of family life become intimately tied up with rooms, windows, and gardens. Each new family creates a new layer of meaning, resulting in a palimpsest of family memory. These houses function quite literally as memory palaces, as a walk through a childhood home will show; each room conjures up images of past events. Over time, these personal memories become woven together with the cultural memory of countries and generations. The importance of the home is a central theme in art, and several contemporary artists have a special interest in the relationship between memory and the home. This paper analyses three case studies in order to get a deeper understanding of the ways in which the home functions and feels like a memory palace, both on an individual and on a collective, cultural level. Close reading of the artworks is performed on the theoretical intersection between Art History and Cultural Memory Studies. The first case study concerns works from the exhibition Mnemosyne by the artist duo Anne and Patrick Poirier. These works combine interests in architecture, archaeology, and psychology. Models of cities and fantastical architectural designs resemble physical structures (such as the brain), architectural metaphors used in representing the concept of memory (such as the memory palace), and archaeological remains, essential to our shared cultural memories. Secondly, works by Do Ho Suh will help us understand the relationship between the home and memory on a far more personal level; outlines of rooms from his former homes, made of colourful, transparent fabric and combined into new structures, provide an insight into the way these spaces retain individual memories. The spaces have been emptied out, and only the husks remain. Although the remnants of walls, light switches, doors, electricity outlets, etc. are standard, mass-produced elements found in many homes and devoid of inherent meaning, together they remind us of the emotional significance attached to the muscle memory of spaces we once inhabited. The third case study concerns an exhibition in a house put up for sale on the Dutch real estate website Funda. The house was built in 1933 by a Jewish family fleeing from Germany, and the father and son were later deported and killed. The artists Anne van As and CA Wertheim have used the history and memories of the house as a starting point for an exhibition called (T)huis, a combination of the Dutch words for home and house. This case study illustrates the way houses become containers of memories; each new family ‘resets’ the meaning of a house, but traces of earlier memories remain. The exhibition allows us to explore the transition of individual memories into shared cultural memory, in this case of WWII. Taken together, the analyses provide a deeper understanding of different facets of the relationship between the home and memory, both individual and collective, and the ways in which art can represent these.

Keywords: Anne and Patrick Poirier, cultural memory, Do Ho Suh, home, memory palace

Procedia PDF Downloads 136
274 Nuancing the Indentured Migration in Amitav Ghosh's Sea of Poppies

Authors: Murari Prasad

Abstract:

This paper is motivated by the implications of indentured migration depicted in Amitav Ghosh’s critically acclaimed novel, Sea of Poppies (2008). Ghosh’s perspective on the experiences of North Indian indentured labourers moving from their homeland to a distant and unknown location across the seas suggests a radical attitudinal change among the migrants on board the Ibis, a schooner chartered to carry the recruits from Calcutta to Mauritius in the late 1830s. The novel unfolds the life-altering trauma of the bonded servants, including their efforts to maintain a sense of self while negotiating significant social and cultural transformations during the voyage which leads to the breakdown of familiar life-worlds. Equally, the migrants are introduced to an alternative network of relationships to ensure their survival away from land. They relinquish their entrenched beliefs and prejudices and commit themselves to a new brotherhood formed by ‘ship siblings.’ With the official abolition of direct slavery in 1833, the supply of cheap labour to the sugar plantation in British colonies as far-flung as Mauritius and Fiji to East Africa and the Caribbean sharply declined. Around the same time, China’s attempt to prohibit the illegal importation of opium from British India into China threatened the lucrative opium trade. To run the ever-profitable plantation colonies with cheap labour, Indian peasants, wrenched from their village economies, were indentured to plantations as girmitiyas (vernacularized from ‘agreement’) by the colonial government using the ploy of an optional form of recruitment. After the British conquest of the Isle of France in 1810, Mauritius became Britain’s premier sugar colony bringing waves of Indian immigrants to the island. In the articulations of their subjectivities one notices how the recruits cope with the alienating drudgery of indenture, mitigate the hardships of the voyage and forge new ties with pragmatic acts of cultural syncretism in a forward-looking autonomous community of ‘ship-siblings’ following the fracture of traditional identities. This paper tests the hypothesis that Ghosh envisions a kind of futuristic/utopian political collectivity in a hierarchically rigid, racially segregated and identity-obsessed world. In order to ground the claim and frame the complex representations of alliance and love across the boundaries of caste, religion, gender and nation, the essential methodology here is a close textual analysis of the novel. This methodology will be geared to explicate the utopian futurity that the novel gestures towards by underlining new regulations of life during voyage and dissolution of multiple differences among the indentured migrants on board the Ibis.

Keywords: indenture, colonial, opium, sugar plantation

Procedia PDF Downloads 374
273 Mixed-Methods Analyses of Subjective Strategies of Most Unlikely but Successful Transitions from Social Benefits to Work

Authors: Hirseland Andreas, Kerschbaumer Lukas

Abstract:

In the case of Germany, there are about one million long-term unemployed – a figure that did not vary much during the past years. These long-term unemployed did not benefit from the prospering labor market while most short-term unemployed did. Instead, they are continuously dependent on welfare and sometimes precarious short-term employment, experiencing work poverty. Long-term unemployment thus turns into a main obstacle to become employed again, especially if it is accompanied by other impediments such as low-level education (school/vocational), poor health (especially chronical illness), advanced age (older than fifty), immigrant status, motherhood or engagement in care for other relatives. As can be shown by this current research project, in these cases the chance to regain employment decreases to near nil. Almost two-thirds of all welfare recipients have multiple impediments which hinder a successful transition from welfare back to sustainable and sufficient employment. Prospective employers are unlikely to hire long-term unemployed with additional impediments because they evaluate potential employees on their negative signaling (e.g. low-level education) and the implicit assumption of unproductiveness (e.g. poor health, age). Some findings of the panel survey “Labor market and social security” (PASS) carried out by the Institute of Employment Research (the research institute of the German Federal Labor Agency) spread a ray of hope, showing that unlikely does not necessarily mean impossible. The presentation reports on current research on these very scarce “success stories” of unlikely transitions from long-term unemployment to work and how these cases were able to perform this switch against all odds. The study is based on a mixed-method design. Within the panel survey (~15,000 respondents in ~10,000 households), only 66 cases of such unlikely transitions were observed. These cases have been explored by qualitative inquiry – in depth-interviews and qualitative network techniques. There is strong evidence that sustainable transitions are influenced by certain biographical resources like habits of network use, a set of informal skills and particularly a resilient way of dealing with obstacles, combined with contextual factors rather than by job-placement procedures promoted by Job-Centers according to activation rules or by following formal paths of application. On the employer’s side small and medium-sized enterprises are often found to give job opportunities to a wider variety of applicants, often based on a slow but steadily increasing relationship leading to employment. According to these results it is possible to show and discuss some limitations of (German) activation policies targeting the labor market and their impact on welfare dependency and long-term unemployment. Based on these findings, indications for more supportive small-scale measures in the field of labor-market policies are suggested to help long-term unemployed with multiple impediments to overcome their situation (e.g. organizing small-scale-structures and low-threshold services to encounter possible employers on a more informal basis like “meet and greet”).

Keywords: against-all-odds, mixed-methods, Welfare State, long-term unemployment

Procedia PDF Downloads 343
272 Task Validity in Neuroimaging Studies: Perspectives from Applied Linguistics

Authors: L. Freeborn

Abstract:

Recent years have seen an increasing number of neuroimaging studies related to language learning as imaging techniques such as fMRI and EEG have become more widely accessible to researchers. By using a variety of structural and functional neuroimaging techniques, these studies have already made considerable progress in terms of our understanding of neural networks and processing related to first and second language acquisition. However, the methodological designs employed in neuroimaging studies to test language learning have been questioned by applied linguists working within the field of second language acquisition (SLA). One of the major criticisms is that tasks designed to measure language learning gains rarely have a communicative function, and seldom assess learners’ ability to use the language in authentic situations. This brings the validity of many neuroimaging tasks into question. The fundamental reason why people learn a language is to communicate, and it is well-known that both first and second language proficiency are developed through meaningful social interaction. With this in mind, the SLA field is in agreement that second language acquisition and proficiency should be measured through learners’ ability to communicate in authentic real-life situations. Whilst authenticity is not always possible to achieve in a classroom environment, the importance of task authenticity should be reflected in the design of language assessments, teaching materials, and curricula. Tasks that bear little relation to how language is used in real-life situations can be considered to lack construct validity. This paper first describes the typical tasks used in neuroimaging studies to measure language gains and proficiency, then analyses to what extent these tasks can validly assess these constructs.

Keywords: neuroimaging studies, research design, second language acquisition, task validity

Procedia PDF Downloads 111
271 Automatic Classification of Lung Diseases from CT Images

Authors: Abobaker Mohammed Qasem Farhan, Shangming Yang, Mohammed Al-Nehari

Abstract:

Pneumonia is a kind of lung disease that creates congestion in the chest. Such pneumonic conditions lead to loss of life of the severity of high congestion. Pneumonic lung disease is caused by viral pneumonia, bacterial pneumonia, or Covidi-19 induced pneumonia. The early prediction and classification of such lung diseases help to reduce the mortality rate. We propose the automatic Computer-Aided Diagnosis (CAD) system in this paper using the deep learning approach. The proposed CAD system takes input from raw computerized tomography (CT) scans of the patient's chest and automatically predicts disease classification. We designed the Hybrid Deep Learning Algorithm (HDLA) to improve accuracy and reduce processing requirements. The raw CT scans have pre-processed first to enhance their quality for further analysis. We then applied a hybrid model that consists of automatic feature extraction and classification. We propose the robust 2D Convolutional Neural Network (CNN) model to extract the automatic features from the pre-processed CT image. This CNN model assures feature learning with extremely effective 1D feature extraction for each input CT image. The outcome of the 2D CNN model is then normalized using the Min-Max technique. The second step of the proposed hybrid model is related to training and classification using different classifiers. The simulation outcomes using the publically available dataset prove the robustness and efficiency of the proposed model compared to state-of-art algorithms.

Keywords: CT scan, Covid-19, deep learning, image processing, lung disease classification

Procedia PDF Downloads 122
270 Text Emotion Recognition by Multi-Head Attention based Bidirectional LSTM Utilizing Multi-Level Classification

Authors: Vishwanath Pethri Kamath, Jayantha Gowda Sarapanahalli, Vishal Mishra, Siddhesh Balwant Bandgar

Abstract:

Recognition of emotional information is essential in any form of communication. Growing HCI (Human-Computer Interaction) in recent times indicates the importance of understanding of emotions expressed and becomes crucial for improving the system or the interaction itself. In this research work, textual data for emotion recognition is used. The text being the least expressive amongst the multimodal resources poses various challenges such as contextual information and also sequential nature of the language construction. In this research work, the proposal is made for a neural architecture to resolve not less than 8 emotions from textual data sources derived from multiple datasets using google pre-trained word2vec word embeddings and a Multi-head attention-based bidirectional LSTM model with a one-vs-all Multi-Level Classification. The emotions targeted in this research are Anger, Disgust, Fear, Guilt, Joy, Sadness, Shame, and Surprise. Textual data from multiple datasets were used for this research work such as ISEAR, Go Emotions, Affect datasets for creating the emotions’ dataset. Data samples overlap or conflicts were considered with careful preprocessing. Our results show a significant improvement with the modeling architecture and as good as 10 points improvement in recognizing some emotions.

Keywords: text emotion recognition, bidirectional LSTM, multi-head attention, multi-level classification, google word2vec word embeddings

Procedia PDF Downloads 151
269 A Comparative Study on Deep Learning Models for Pneumonia Detection

Authors: Hichem Sassi

Abstract:

Pneumonia, being a respiratory infection, has garnered global attention due to its rapid transmission and relatively high mortality rates. Timely detection and treatment play a crucial role in significantly reducing mortality associated with pneumonia. Presently, X-ray diagnosis stands out as a reasonably effective method. However, the manual scrutiny of a patient's X-ray chest radiograph by a proficient practitioner usually requires 5 to 15 minutes. In situations where cases are concentrated, this places immense pressure on clinicians for timely diagnosis. Relying solely on the visual acumen of imaging doctors proves to be inefficient, particularly given the low speed of manual analysis. Therefore, the integration of artificial intelligence into the clinical image diagnosis of pneumonia becomes imperative. Additionally, AI recognition is notably rapid, with convolutional neural networks (CNNs) demonstrating superior performance compared to human counterparts in image identification tasks. To conduct our study, we utilized a dataset comprising chest X-ray images obtained from Kaggle, encompassing a total of 5216 training images and 624 test images, categorized into two classes: normal and pneumonia. Employing five mainstream network algorithms, we undertook a comprehensive analysis to classify these diseases within the dataset, subsequently comparing the results. The integration of artificial intelligence, particularly through improved network architectures, stands as a transformative step towards more efficient and accurate clinical diagnoses across various medical domains.

Keywords: deep learning, computer vision, pneumonia, models, comparative study

Procedia PDF Downloads 35
268 Times2D: A Time-Frequency Method for Time Series Forecasting

Authors: Reza Nematirad, Anil Pahwa, Balasubramaniam Natarajan

Abstract:

Time series data consist of successive data points collected over a period of time. Accurate prediction of future values is essential for informed decision-making in several real-world applications, including electricity load demand forecasting, lifetime estimation of industrial machinery, traffic planning, weather prediction, and the stock market. Due to their critical relevance and wide application, there has been considerable interest in time series forecasting in recent years. However, the proliferation of sensors and IoT devices, real-time monitoring systems, and high-frequency trading data introduce significant intricate temporal variations, rapid changes, noise, and non-linearities, making time series forecasting more challenging. Classical methods such as Autoregressive integrated moving average (ARIMA) and Exponential Smoothing aim to extract pre-defined temporal variations, such as trends and seasonality. While these methods are effective for capturing well-defined seasonal patterns and trends, they often struggle with more complex, non-linear patterns present in real-world time series data. In recent years, deep learning has made significant contributions to time series forecasting. Recurrent Neural Networks (RNNs) and their variants, such as Long short-term memory (LSTMs) and Gated Recurrent Units (GRUs), have been widely adopted for modeling sequential data. However, they often suffer from the locality, making it difficult to capture local trends and rapid fluctuations. Convolutional Neural Networks (CNNs), particularly Temporal Convolutional Networks (TCNs), leverage convolutional layers to capture temporal dependencies by applying convolutional filters along the temporal dimension. Despite their advantages, TCNs struggle with capturing relationships between distant time points due to the locality of one-dimensional convolution kernels. Transformers have revolutionized time series forecasting with their powerful attention mechanisms, effectively capturing long-term dependencies and relationships between distant time points. However, the attention mechanism may struggle to discern dependencies directly from scattered time points due to intricate temporal patterns. Lastly, Multi-Layer Perceptrons (MLPs) have also been employed, with models like N-BEATS and LightTS demonstrating success. Despite this, MLPs often face high volatility and computational complexity challenges in long-horizon forecasting. To address intricate temporal variations in time series data, this study introduces Times2D, a novel framework that parallelly integrates 2D spectrogram and derivative heatmap techniques. The spectrogram focuses on the frequency domain, capturing periodicity, while the derivative patterns emphasize the time domain, highlighting sharp fluctuations and turning points. This 2D transformation enables the utilization of powerful computer vision techniques to capture various intricate temporal variations. To evaluate the performance of Times2D, extensive experiments were conducted on standard time series datasets and compared with various state-of-the-art algorithms, including DLinear (2023), TimesNet (2023), Non-stationary Transformer (2022), PatchTST (2023), N-HiTS (2023), Crossformer (2023), MICN (2023), LightTS (2022), FEDformer (2022), FiLM (2022), SCINet (2022a), Autoformer (2021), and Informer (2021) under the same modeling conditions. The initial results demonstrated that Times2D achieves consistent state-of-the-art performance in both short-term and long-term forecasting tasks. Furthermore, the generality of the Times2D framework allows it to be applied to various tasks such as time series imputation, clustering, classification, and anomaly detection, offering potential benefits in any domain that involves sequential data analysis.

Keywords: derivative patterns, spectrogram, time series forecasting, times2D, 2D representation

Procedia PDF Downloads 18
267 Early Depression Detection for Young Adults with a Psychiatric and AI Interdisciplinary Multimodal Framework

Authors: Raymond Xu, Ashley Hua, Andrew Wang, Yuru Lin

Abstract:

During COVID-19, the depression rate has increased dramatically. Young adults are most vulnerable to the mental health effects of the pandemic. Lower-income families have a higher ratio to be diagnosed with depression than the general population, but less access to clinics. This research aims to achieve early depression detection at low cost, large scale, and high accuracy with an interdisciplinary approach by incorporating clinical practices defined by American Psychiatric Association (APA) as well as multimodal AI framework. The proposed approach detected the nine depression symptoms with Natural Language Processing sentiment analysis and a symptom-based Lexicon uniquely designed for young adults. The experiments were conducted on the multimedia survey results from adolescents and young adults and unbiased Twitter communications. The result was further aggregated with the facial emotional cues analyzed by the Convolutional Neural Network on the multimedia survey videos. Five experiments each conducted on 10k data entries reached consistent results with an average accuracy of 88.31%, higher than the existing natural language analysis models. This approach can reach 300+ million daily active Twitter users and is highly accessible by low-income populations to promote early depression detection to raise awareness in adolescents and young adults and reveal complementary cues to assist clinical depression diagnosis.

Keywords: artificial intelligence, COVID-19, depression detection, psychiatric disorder

Procedia PDF Downloads 109