Search results for: vision transformer

Authors: Molakala Mourya Vardhan Reddy, Kandimala Revanth, Koduru Sumanth, Beena B. M.

Abstract:

Emergency response and recovery operations are severely hampered by natural catastrophes, especially floods. Understanding post-flood scenarios is essential to disaster management because it facilitates quick evaluation and decision-making. To this end, we introduce FloodNet, a brand-new high-resolution aerial picture collection created especially for comprehending post-flood scenes. A varied collection of excellent aerial photos taken during and after flood occurrences make up FloodNet, which offers comprehensive representations of flooded landscapes, damaged infrastructure, and changed topographies. The dataset provides a thorough resource for training and assessing computer vision models designed to handle the complexity of post-flood scenarios, including a variety of environmental conditions and geographic regions. Pixel-level semantic segmentation masks are used to label the pictures in FloodNet, allowing for a more detailed examination of flood-related characteristics, including debris, water bodies, and damaged structures. Furthermore, temporal and positional metadata improve the dataset's usefulness for longitudinal research and spatiotemporal analysis. For activities like flood extent mapping, damage assessment, and infrastructure recovery projection, we provide baseline standards and evaluation metrics to promote research and development in the field of post-flood scene comprehension. By integrating FloodNet into machine learning pipelines, it will be easier to create reliable algorithms that will help politicians, urban planners, and first responders make choices both before and after floods. The goal of the FloodNet dataset is to support advances in computer vision, remote sensing, and disaster response technologies by providing a useful resource for researchers. FloodNet helps to create creative solutions for boosting communities' resilience in the face of natural catastrophes by tackling the particular problems presented by post-flood situations.

Keywords: image classification, segmentation, computer vision, nature disaster, unmanned arial vehicle(UAV), machine learning.

Procedia PDF Downloads 55

Authors: Hebah M. Musalem, Jeylan El-Mansoury, Lin M. Tuleimat, Selwa Alhazza, Abdul-Aziz A. Al Zoba

Abstract:

Background: Amblyopia is a condition that affects the visual system triggering a decrease in visual acuity without a known underlying pathology. It is due to abnormal vision development in childhood or infancy. Most importantly, vision loss is preventable or reversible with the right kind of intervention in most of the cases. Strabismus, sensory defects, and anisometropia are all well-known causes of amblyopia. However, ocular misalignment in Strabismus is considered the most common form of amblyopia worldwide. The risk of developing amblyopia increases in premature children, developmentally delayed or children who had brain lesions affecting the visual pathway. The prevalence of amblyopia varies between 2 to 5 % in the world according to the literature. Objective: To determine the different causes of Amblyopia in pediatric patients seen in ophthalmology clinic of a tertiary care center, i.e. King Faisal Specialist Hospital and Research Center (KFSH&RC). Methods: This is a hospital based, random retrospective, based on reviewing patient’s files in the Ophthalmology Department of KFSH&RC in Riyadh city, Kingdom of Saudi Arabia. Inclusion criteria: amblyopic pediatric patients who attended the clinic from 2015 to 2016, who are between 6 months and 18 years old. Exclusion Criteria: patients above 18 years of age and any patient who is uncooperative to obtain an accurate vision or a proper refraction. Detailed ocular and medical history are recorded. The examination protocol includes a full ocular exam, full cycloplegic refraction, visual acuity measurement, ocular motility and strabismus evaluation. All data were organized in tables and graphs and analyzed by statistician. Results: Our preliminary results will be discussed on spot by our corresponding author. Conclusions: We focused on this study on utilizing various examination techniques which enhanced our results and highlighted a distinguished correlation between amblyopia and its’ causes. This paper recommendation emphasizes on critical testing protocols to be followed among amblyopic patient, especially in tertiary care centers.

Keywords: amblyopia, amblyopia causes, amblyopia diagnostic criterion, amblyopia prevalence, Saudi Arabia

Procedia PDF Downloads 139

Authors: A. Degale Desta, Cheng Jian

Abstract:

Deep learning application in computer vision is rapidly advancing, giving it the ability to monitor the public and quickly identify potentially anomalous behaviour from crowd scenes. Therefore, the purpose of the current work is to improve the performance of safety of people in crowd events from panic behaviour through introducing the innovative idea of Aggregation of Ensembles (AOE), which makes use of the pre-trained ConvNets and a pool of classifiers to find anomalies in video data with packed scenes. According to the theory of algorithms that applied K-means, KNN, CNN, SVD, and Faster-CNN, YOLOv5 architectures learn different levels of semantic representation from crowd videos; the proposed approach leverages an ensemble of various fine-tuned convolutional neural networks (CNN), allowing for the extraction of enriched feature sets. In addition to the above algorithms, a long short-term memory neural network to forecast future feature values and a handmade feature that takes into consideration the peculiarities of the crowd to understand human behavior. On well-known datasets of panic situations, experiments are run to assess the effectiveness and precision of the suggested method. Results reveal that, compared to state-of-the-art methodologies, the system produces better and more promising results in terms of accuracy and processing speed.

Keywords: action recognition, computer vision, crowd detecting and tracking, deep learning

Procedia PDF Downloads 141

Authors: Zheng Baichuan, Li Shenghui, Li Bingqian, Zhang Ning, Chen Kai

Abstract:

Entity alignment (EA) tasks in knowledge graphs often play a pivotal role in the integration of knowledge graphs, where structural differences often exist between the source and target graphs, such as the presence or absence of attribute information and the types of attribute information (text, timestamps, images, etc.). However, most current research efforts are focused on improving alignment accuracy, often along with an increased reliance on specific structures -a dependency that inevitably diminishes their practical value and causes difficulties when facing knowledge graph alignment tasks with varying structures. Therefore, we propose a universal knowledge graph alignment approach that only utilizes the common basic structures shared by knowledge graphs. We have demonstrated through experiments that our method achieves state-of-the-art performance in fair comparisons.

Keywords: knowledge graph, entity alignment, transformer, deep learning

Procedia PDF Downloads 27

Authors: Noga Bregman

Abstract:

Accurate and efficient earthquake detection and phase picking are crucial for seismic hazard assessment and emergency response. This study introduces EQMamba, a deep-learning method that combines the strengths of the Earthquake Transformer and the Mamba model for simultaneous earthquake detection and phase picking. EQMamba leverages the computational efficiency of Mamba layers to process longer seismic sequences while maintaining a manageable model size. The proposed architecture integrates convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM) networks, and Mamba blocks. The model employs an encoder composed of convolutional layers and max pooling operations, followed by residual CNN blocks for feature extraction. Mamba blocks are applied to the outputs of BiLSTM blocks, efficiently capturing long-range dependencies in seismic data. Separate decoders are used for earthquake detection, P-wave picking, and S-wave picking. We trained and evaluated EQMamba using a subset of the STEAD dataset, a comprehensive collection of labeled seismic waveforms. The model was trained using a weighted combination of binary cross-entropy loss functions for each task, with the Adam optimizer and a scheduled learning rate. Data augmentation techniques were employed to enhance the model's robustness. Performance comparisons were conducted between EQMamba and the EQTransformer over 20 epochs on this modest-sized STEAD subset. Results demonstrate that EQMamba achieves superior performance, with higher F1 scores and faster convergence compared to EQTransformer. EQMamba reached F1 scores of 0.8 by epoch 5 and maintained higher scores throughout training. The model also exhibited more stable validation performance, indicating good generalization capabilities. While both models showed lower accuracy in phase-picking tasks compared to detection, EQMamba's overall performance suggests significant potential for improving seismic data analysis. The rapid convergence and superior F1 scores of EQMamba, even on a modest-sized dataset, indicate promising scalability for larger datasets. This study contributes to the field of earthquake engineering by presenting a computationally efficient and accurate method for simultaneous earthquake detection and phase picking. Future work will focus on incorporating Mamba layers into the P and S pickers and further optimizing the architecture for seismic data specifics. The EQMamba method holds the potential for enhancing real-time earthquake monitoring systems and improving our understanding of seismic events.

Keywords: earthquake, detection, phase picking, s waves, p waves, transformer, deep learning, seismic waves

Procedia PDF Downloads 14

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 87

Authors: Alexandre Faure, Yoko Mizokami, éRic Dinet

Abstract:

These recent years, the potential of light in assisting visually impaired people in their indoor mobility has been demonstrated by different studies. Implementing smart lighting systems for selective visual enhancement, especially designed for low-vision people, is an approach that breaks with the existing visual aids. The appearance of the surface of an object is significantly influenced by the lighting conditions and the constituent materials of the objects. Appearance of objects may appear to be different from expectation. Therefore, lighting conditions lead to an important part of accurate material recognition. The main objective of this work was to investigate the effect of the spatial distribution of light on object identification in the context of low vision. The purpose was to determine whether and what specific lighting approaches should be preferred for visually impaired people. A psychophysical experiment was designed to study the ability of individuals to identify the smallest cube of a pair under different lighting diffusion conditions. Participants were divided into two distinct groups: a reference group of observers with normal or corrected-to-normal visual acuity and a test group, in which observers were required to wear visual impairment simulation glasses. All participants were presented with pairs of cubes in a "miniature room" and were instructed to estimate the relative size of the two cubes. The miniature room replicates real-life settings, adorned with decorations and separated from external light sources by black curtains. The correlated color temperature was set to 6000 K, and the horizontal illuminance at the object level at approximately 240 lux. The objects presented for comparison consisted of 11 white cubes and 11 black cubes of different sizes manufactured with a 3D printer. Participants were seated 60 cm away from the objects. Two different levels of light diffuseness were implemented. After receiving instructions, participants were asked to judge whether the two presented cubes were the same size or if one was smaller. They provided one of five possible answers: "Left one is smaller," "Left one is smaller but unsure," "Same size," "Right one is smaller," or "Right one is smaller but unsure.". The method of constant stimuli was used, presenting stimulus pairs in a random order to prevent learning and expectation biases. Each pair consisted of a comparison stimulus and a reference cube. A psychometric function was constructed to link stimulus value with the frequency of correct detection, aiming to determine the 50% correct detection threshold. Collected data were analyzed through graphs illustrating participants' responses to stimuli, with accuracy increasing as the size difference between cubes grew. Statistical analyses, including 2-way ANOVA tests, showed that light diffuseness had no significant impact on the difference threshold, whereas object color had a significant influence in low vision scenarios. The first results and trends derived from this pilot experiment clearly and strongly suggest that future investigations could explore extreme diffusion conditions to comprehensively assess the impact of diffusion on object identification. For example, the first findings related to light diffuseness may be attributed to the range of manipulation, emphasizing the need to explore how other lighting-related factors interact with diffuseness.

Keywords: Lighting, Low Vision, Visual Aid, Object Identification, Psychophysical Experiment

Procedia PDF Downloads 47

Authors: Delphine Vendryes, Dushyanth Sekhar, Baojia Tong, Matthew Theisen, Chester Curme

Abstract:

Much of the data that inform the decisions of governments, corporations and individuals are harvested from unstructured documents. Data extraction is defined here as a process that turns non-machine-readable information into a machine-readable format that can be stored, for instance, in a database. In financial services, introducing more automation in data extraction pipelines is a major challenge. Information sought by financial data consumers is often buried within vast bodies of unstructured documents, which have historically required thorough manual extraction. Automated solutions provide faster access to non-machine-readable datasets, in a context where untimely information quickly becomes irrelevant. Data quality standards cannot be compromised, so automation requires high data integrity. This multifaceted task is broken down into smaller steps: ingestion, table parsing (detection and structure recognition), text analysis (entity detection and disambiguation), schema-based record extraction, user feedback incorporation. Selected intermediary steps are phrased as machine learning problems. Solutions leveraging cutting-edge approaches from the fields of computer vision (e.g. table detection) and natural language processing (e.g. entity detection and disambiguation) are proposed.

Keywords: computer vision, entity recognition, finance, information retrieval, machine learning, natural language processing

Procedia PDF Downloads 93

Authors: A. M. Al-Sabagh, M. A. Deyab, M. N. Kroush

Abstract:

The present study focused on critical and detailed approach for using aluminum electrode as impressed current anode for cathodic protection of X65 steel tank containing fresh water. The impressed current design calculation showed 0.6 A of current demand and voltage of 0.33 V required to adequately protect the X65 steel tank with internal surface area of 421 m². We used here one transformer rectifier with current and voltage output of 25 A and 25 V, respectively. The data showed that the potentials ranged from -0.474 to -0.509 V (vs. Cu/CuSO₄), prior to the application of cathodic protection. When the potential was measured 1 h after the application of cathodic protection, the potential values showed considerable shift within protection range (-0.950 V vs. Cu/CuSO₄). The results confirmed that aluminum anode can be used in freshwater applications with high efficiency (current capacity) and low consumption rate.

Keywords: cathodic protection, aluminum, steel, fresh water

Procedia PDF Downloads 133

Authors: Sana Abusin

Abstract:

Food security in Qatar is a research priority of Qatar University (2021-2025) and all national strategies, including the Qatar National Vision 2030 and food security strategy (2018-2023). Achieving food security requires three actions: 1) transforming surplus food to those who are insecure; 2) reducing food loss and waste by recycling food into valuable resources such as compost (“green fertilizer”) that can be used in growing food; and, finally, 3) establishing strong enforcement agencies to protect consumers from outdated food and promote healthy food. Currently, these objectives are approached separately and not in a sustainable fashion. Food security in Qatar is a research priority of Qatar University (2021-2025) and all national strategies, including the Qatar National Vision 2030 and food security strategy (2018-2023). The study aims to develop an innovative mobile application that supports a sustainable solution to food insecurity and food waste in Qatar. The application will provide a common solution for many different users. For producers, it will facilitate easy disposal of excess food. For charities, it will notify them about surplus food ready for redistribution. The application will also benefit the second layer of end-users in the form of food recycling companies, who will receive information about available food waste that is unable to be consumed. We will use self-exoplanetary diagrams and digital pictures to show all the steps to the final stage. The aim is to motivate the young generation toward innovation and creation, and to encourage public-private collaboration in this sector.

Keywords: food security, innovative technology, sustainability, food waste, Qatar

Procedia PDF Downloads 98

Authors: Raz Flieshman, Adam Michael Altenbuchner, Jörg Krüger

Abstract:

This paper presents an approach to optimizing the Fused Deposition Modeling (FDM) 3D printing process through a preprocess calibration routine of printing parameters. The core of this method involves the use of a low-cost thermal sensor capable of measuring tempera-tures within the range of -20 to 500 degrees Celsius for detailed process observation. The calibration process is conducted by printing a predetermined path while varying the process parameters through machine instructions (g-code). This enables the extraction of critical thermal, dimensional, and surface properties along the printed path. The calibration routine utilizes computer vision models to extract features and metrics from the thermal images, in-cluding temperature distribution, layer adhesion quality, surface roughness, and dimension-al accuracy and consistency. These extracted properties are then analyzed to optimize the process parameters to achieve the desired qualities of the printed material. A significant benefit of this calibration method is its potential to create printing parameter profiles for new polymer and composite materials, thereby enhancing the versatility and application range of FDM 3D printing. The proposed method demonstrates significant potential in enhancing the precision and reliability of FDM 3D printing, making it a valuable contribution to the field of additive manufacturing.

Keywords: FDM 3D printing, preprocess calibration, thermal sensor, process optimization, additive manufacturing, computer vision, material profiles

Procedia PDF Downloads 12

Authors: Sanjeev Gupta, Ankit Prabhakar, K. Rajkumar Singh

Abstract:

Sunni Dam Hydroelectric Project (382 MW) is a run of river type development with an underground powerhouse, proposed to harness the hydel potential of river Satluj in Himachal Pradesh, India. The project is located in the inner lesser Himalaya between Dhauladhar Range in the south and the higher Himalaya in the north. The project comprises two large underground caverns, a Powerhouse cavern (171m long, 22.5m wide and 51.2m high) and another transformer hall cavern (175m long, 18.7m wide and 27m high) and the rock pillar between the two caverns is 50m. The highly jointed, fractured, anisotropic rock mass is a key challenge in Himalayan geology for an underground structure. The concern for the stability of rock mass increases when weak/shear zones are encountered in the underground structure. In the Sunni Dam project, 1.7m to 2m thick weak/shear zone comprising of deformed, weak material with gauge has been encountered in powerhouse cavern at 70m having dip direction 325 degree and dip amount 38 degree which also intersects transformer hall at initial reach. The rock encountered in the powerhouse area is moderate to highly jointed, pink quartz arenite belonging to the Khaira Formation, a transition zone comprising of alternate grey, pink & white quartz arenite and shale sequence and dolomite at higher reaches. The rock mass is intersected by mainly 3 joint sets excluding bedding joints and a few random joints. The rock class in powerhouse mainly varies from poor class (class IV) to lower order fair class (class III) and in some reaches, very poor rock mass has also been encountered. To study the stability of the underground structure in weak/shear rock mass, a 3D numerical model analysis has been carried out using RS3 software. Field studies have been interpreted and analysed to derive Bieniawski’s RMR, Barton’s “Q” class and Geological Strength Index (GSI). The various material parameters, in-situ characteristics have been determined based on tests conducted by Central Soil and Materials Research Station, New Delhi. The behaviour of the cavern has been studied by assessing the displacement contours, major and minor principal stresses and plastic zones for different stage excavation sequences. For optimisation of the support system, the stability of the powerhouse cavern with different powerhouse orientations has also been studied. The numerical modeling results indicate that cavern will not likely face stress governed by structural instability with the support system to be applied to the crown and side walls.

Keywords: 3D analysis, Himalayan geology, shear zone, underground power house

Procedia PDF Downloads 68

Authors: Shuruq Ghazi Nahhas

Abstract:

Islamic art is a decorative art that depends on repeating motifs in various shapes to cover different surfaces. Each motif has its own characteristics and style that may reflect different Islamic periods, such as Umayyad, Abbasid, Fatimid, Seljuk, Nasrid, Ottoman, and Safavid. These periods were the most powerful periods which played an important role in developing the Islamic motifs. Most of these motifs of the Islamic heritage were not used in new applications. This research focused on reviving the vegetal Islamic motifs found on Islamic heritage and redesign them in a new format to decorate various products, including scarfs, cushions, coasters, wallpaper, wall art, and boxes. The scarf is chosen as one element of these decorative products because it is used as accessories to add aesthetic value to fashion. A descriptive-analytical method is used for this research. The process started with extracting and analyzing the original motifs. Then, creating the new motifs by simplifying, deleting, or adding elements based on the original structure. Then, creating repeated patterns and applying them to decorative products. The findings of this research indicated: repeating patterns based on different structures creates unlimited patterns. Also, changing the elements of the motifs of a pattern adds new characteristics to the pattern. Also, creating frames using elements from the repeated motifs adds aesthetic and contemporary value to decorative products. Finally, using various methods of combining colors creates unlimited variations of each pattern. At the end, reviving the Islamic motifs in contemporary vision enriches decorative products with aesthetic, artistic, and historical values of different Islamic periods. This makes the decorative products valuable that adds uniqueness to their surroundings.

Keywords: Islamic motifs, contemporary patterns, scarfs, decorative products

Procedia PDF Downloads 146

Authors: Colin Schmidt, Adrien Lecossier, Pascal Crubleau, Simon Richir

Abstract:

Introduction:The widespread availability of artificial intelligence, exemplified by Generative Pre-trained Transformers (GPT) relying on large language models (LLM), has caused a seismic shift in the realm of knowledge. Everyone now has the capacity to swiftly learn how these models can either serve them well or not. Today, conversational AI like ChatGPT is grounded in neural transformer models, a significant advance in natural language processing facilitated by the emergence of renowned LLMs constructed using neural transformer architecture. Inventiveness of an LLM : OpenAI's GPT-3 stands as a premier LLM, capable of handling a broad spectrum of natural language processing tasks without requiring fine-tuning, reliably producing text that reads as if authored by humans. However, even with an understanding of how LLMs respond to questions asked, there may be lurking behind OpenAI’s seemingly endless responses an inventive model yet to be uncovered. There may be some unforeseen reasoning emerging from the interconnection of neural networks here. Just as a Soviet researcher in the 1940s questioned the existence of Common factors in inventions, enabling an Under standing of how and according to what principles humans create them, it is equally legitimate today to explore whether solutions provided by LLMs to complex problems also share common denominators. Theory of Inventive Problem Solving (TRIZ) : We will revisit some fundamentals of TRIZ and how Genrich ALTSHULLER was inspired by the idea that inventions and innovations are essential means to solve societal problems. It's crucial to note that traditional problem-solving methods often fall short in discovering innovative solutions. The design team is frequently hampered by psychological barriers stemming from confinement within a highly specialized knowledge domain that is difficult to question. We presume ChatGPT Utilizes TRIZ 40. Hence, the objective of this research is to decipher the inventive model of LLMs, particularly that of ChatGPT, through a comparative study. This will enhance the efficiency of sustainable innovation processes and shed light on how the construction of a solution to a complex problem was devised. Description of the Experimental Protocol : To confirm or reject our main hypothesis that is to determine whether ChatGPT uses TRIZ, we will follow a stringent protocol that we will detail, drawing on insights from a panel of two TRIZ experts. Conclusion and Future Directions : In this endeavor, we sought to comprehend how an LLM like GPT addresses complex challenges. Our goal was to analyze the inventive model of responses provided by an LLM, specifically ChatGPT, by comparing it to an existing standard model: TRIZ 40. Of course, problem solving is our main focus in our endeavours.

Keywords: artificial intelligence, Triz, ChatGPT, inventiveness, problem-solving

Procedia PDF Downloads 44

Authors: Iqraq Kamal, Akmal Razif, Sivadas Chandra Sekaran, Ahmad Syazwan Hisaburi

Abstract:

Painting plays a crucial role in the aerospace manufacturing industry, serving both protective and cosmetic purposes for components. However, the traditional manual painting method is time-consuming and labor-intensive, posing challenges for the sector in achieving higher efficiency. Additionally, the current automated robot path planning has been a bottleneck for spray painting processes, as typical manual teaching methods are time-consuming, error-prone, and skill-dependent. Therefore, it is essential to develop automated tool path planning methods to replace manual ones, reducing costs and improving product quality. Focusing on flat panel painting in aerospace manufacturing, this study aims to address issues related to unreliable part identification techniques caused by the high-mixture, low-volume nature of the industry. The proposed solution involves using a spray gun and a UR10 robotic arm with a vision system that utilizes one-shot object detection (OS2D) to identify parts accurately. Additionally, the research optimizes path planning by concentrating on the region of interest—specifically, the identified part, rather than uniformly covering the entire painting tray.

Keywords: aerospace manufacturing, one-shot object detection, automated spray painting, vision-based path optimization, deep learning, automation, robotic arm

Procedia PDF Downloads 63

Authors: Maamar Latroch, Mohamed Bourahla

Abstract:

Forward topology used in switch mode power supply (SMPS) is one of the most famous configuration feeding DC systems such as telecommunication systems and other specific applications where the galvanic isolation is required. This configuration benefits of the high frequency feature of the transformer to provide a small size and light weight of the over all system. However, the stresses existing on the power switch during an ON/OFF commutation limit the transmitted power to the DC load. This paper investigates the main causes of the stresses in voltage existing during a commutation cycle and suggest a low cost solution that eliminates the overvoltage. As a result, this configuration will yield the possibility of the use of this configuration in higher power applications. Simulation results will show the efficiency of the presented method.

Keywords: switch mode power supply, forward topology, resonance topology, high frequency commutation

Procedia PDF Downloads 422

Authors: George Zhou, Yunchan Chen, Candace Chien

Abstract:

Kidney replacement therapy is the current standard of care for end-stage renal diseases. In-center or home hemodialysis remains an integral component of the therapeutic regimen. Arteriovenous fistulas (AVF) make up the vascular circuit through which blood is filtered and returned. Naturally, AVF patency determines whether adequate clearance and filtration can be achieved and directly influences clinical outcomes. Our aim was to build a deep learning model for automated AVF stenosis screening based on the sound of blood flow through the AVF. A total of 311 patients with AVF were enrolled in this study. Blood flow sounds were collected using a digital stethoscope. For each patient, blood flow sounds were collected at 6 different locations along the patient’s AVF. The 6 locations are artery, anastomosis, distal vein, middle vein, proximal vein, and venous arch. A total of 1866 sounds were collected. The blood flow sounds are labeled as “patent” (normal) or “stenotic” (abnormal). The labels are validated from concurrent ultrasound. Our dataset included 1527 “patent” and 339 “stenotic” sounds. We show that blood flow sounds vary significantly along the AVF. For example, the blood flow sound is loudest at the anastomosis site and softest at the cephalic arch. Contextualizing the sound with location metadata significantly improves classification performance. How to encode and incorporate categorical metadata is an active area of research1. Herein, we study ordinal (i.e., integer) encoding schemes. The numerical representation is concatenated to the flattened feature vector. We train a vision transformer (ViT) on spectrogram image representations of the sound and demonstrate that using scalar multiples of our integer encodings improves classification performance. Models are evaluated using a 10-fold cross-validation procedure. The baseline performance of our ViT without any location metadata achieves an AuROC and AuPRC of 0.68 ± 0.05 and 0.28 ± 0.09, respectively. Using the following encodings of Artery:0; Arch: 1; Proximal: 2; Middle: 3; Distal 4: Anastomosis: 5, the ViT achieves an AuROC and AuPRC of 0.69 ± 0.06 and 0.30 ± 0.10, respectively. Using the following encodings of Artery:0; Arch: 10; Proximal: 20; Middle: 30; Distal 40: Anastomosis: 50, the ViT achieves an AuROC and AuPRC of 0.74 ± 0.06 and 0.38 ± 0.10, respectively. Using the following encodings of Artery:0; Arch: 100; Proximal: 200; Middle: 300; Distal 400: Anastomosis: 500, the ViT achieves an AuROC and AuPRC of 0.78 ± 0.06 and 0.43 ± 0.11. respectively. Interestingly, we see that using increasing scalar multiples of our integer encoding scheme (i.e., encoding “venous arch” as 1,10,100) results in progressively improved performance. In theory, the integer values do not matter since we are optimizing the same loss function; the model can learn to increase or decrease the weights associated with location encodings and converge on the same solution. However, in the setting of limited data and computation resources, increasing the importance at initialization either leads to faster convergence or helps the model escape a local minimum.

Keywords: arteriovenous fistula, blood flow sounds, metadata encoding, deep learning

Procedia PDF Downloads 65

Authors: Muhammad Bilal Shaikh

Abstract:

Mineral processing plants play a pivotal role in extracting valuable minerals from raw ores, contributing significantly to various industries. However, the optimization of plant throughput remains a complex challenge, necessitating innovative approaches for increased efficiency and productivity. This research paper investigates the application of Multimodal Artificial Intelligence (MAI) techniques to address this challenge, aiming to improve overall plant throughput in mineral processing operations. The integration of multimodal AI leverages a combination of diverse data sources, including sensor data, images, and textual information, to provide a holistic understanding of the complex processes involved in mineral extraction. The paper explores the synergies between various AI modalities, such as machine learning, computer vision, and natural language processing, to create a comprehensive and adaptive system for optimizing mineral processing plants. The primary focus of the research is on developing advanced predictive models that can accurately forecast various parameters affecting plant throughput. Utilizing historical process data, machine learning algorithms are trained to identify patterns, correlations, and dependencies within the intricate network of mineral processing operations. This enables real-time decision-making and process optimization, ultimately leading to enhanced plant throughput. Incorporating computer vision into the multimodal AI framework allows for the analysis of visual data from sensors and cameras positioned throughout the plant. This visual input aids in monitoring equipment conditions, identifying anomalies, and optimizing the flow of raw materials. The combination of machine learning and computer vision enables the creation of predictive maintenance strategies, reducing downtime and improving the overall reliability of mineral processing plants. Furthermore, the integration of natural language processing facilitates the extraction of valuable insights from unstructured textual data, such as maintenance logs, research papers, and operator reports. By understanding and analyzing this textual information, the multimodal AI system can identify trends, potential bottlenecks, and areas for improvement in plant operations. This comprehensive approach enables a more nuanced understanding of the factors influencing throughput and allows for targeted interventions. The research also explores the challenges associated with implementing multimodal AI in mineral processing plants, including data integration, model interpretability, and scalability. Addressing these challenges is crucial for the successful deployment of AI solutions in real-world industrial settings. To validate the effectiveness of the proposed multimodal AI framework, the research conducts case studies in collaboration with mineral processing plants. The results demonstrate tangible improvements in plant throughput, efficiency, and cost-effectiveness. The paper concludes with insights into the broader implications of implementing multimodal AI in mineral processing and its potential to revolutionize the industry by providing a robust, adaptive, and data-driven approach to optimizing plant operations. In summary, this research contributes to the evolving field of mineral processing by showcasing the transformative potential of multimodal artificial intelligence in enhancing plant throughput. The proposed framework offers a holistic solution that integrates machine learning, computer vision, and natural language processing to address the intricacies of mineral extraction processes, paving the way for a more efficient and sustainable future in the mineral processing industry.

Keywords: multimodal AI, computer vision, NLP, mineral processing, mining

Procedia PDF Downloads 53

Authors: M. A. Zaki Ewiss, S. Afifi

Abstract:

Nahda University (NUB) believes that internationalisation of higher educational is able to provide global society with an education that meets current needs and that can respond efficiently to contemporary demands and challenges, which are characterized by globalisation, interdependence, and multiculturalism. In this paper, we will discuss the the challenges of the Egyptian Higher Education system and the future vision to improve this system> In this report, the following issues will be considered: Increasing knowledge on the development of specialized programs of study at the university. Developing international cooperation programs, which focus on the development of the students and staff skills, and providing academic culture and learning opportunities. Increasing the opportunities for student mobility, and research projects for faculty members. Increased opportunities for staff, faculty and students to continue to learn foreign universities, and to benefit from scholarships in various disciplines. Taking the advantage of the educational experience and modern teaching methods; Providing the opportunities to study abroad without increasing the period of time required for graduation, and through greater integration in the curricula and programs; More cultural interaction through student exchanges.Improving and providing job opportunities for graduates through participation in the global labor market. This document sets out NUB strategy to move towards that vision. We are confident that greater explicit differentiation, greater freedom and greater collaboration are the keys to delivering the further improvement in quality we shall need to retain and strengthen our position as one of the world’s leading higher education systems.

Keywords: technology transfer higher education, knowledge transfer, internationalisation, mobility

Procedia PDF Downloads 424

Authors: Xi Liu, Fan Yuan

Abstract:

In the evolving landscape of physics-based machine learning (PBML), particularly within fluid dynamics and its applications in electromechanical engineering, robot vision, and robot learning, achieving precision and alignment with researchers' specific needs presents a formidable challenge. In response, this work proposes a methodology that integrates neural transformation with a modified smoothed particle hydrodynamics model for generating transformed 3D fluid simulations. This approach is useful for nanoscale science, where the unique and complex behaviors of viscoelastic medium demand accurate neurally-transformed simulations for materials understanding and manipulation. In electromechanical engineering, the method enhances the design and functionality of fluid-operated systems, particularly microfluidic devices, contributing to advancements in nanomaterial design, drug delivery systems, and more. The proposed approach also aligns with the principles of PBML, offering advantages such as multi-fluid stylization and consistent particle attribute transfer. This capability is valuable in various fields where the interaction of multiple fluid components is significant. Moreover, the application of neurally-transformed hydrodynamical models extends to manufacturing processes, such as the production of microelectromechanical systems, enhancing efficiency and cost-effectiveness. The system's ability to perform neural transfer on 3D fluid scenes using a deep learning algorithm alongside physical models further adds a layer of flexibility, allowing researchers to tailor simulations to specific needs across scientific and engineering disciplines.

Keywords: physics-based machine learning, robot vision, robot learning, hydrodynamics

Procedia PDF Downloads 48

Authors: Shariful Islam, Sharun Akter Khushbu, S. M. Shaqib, Shahriar Sultan Ramit

Abstract:

In the construction sector, ensuring worker safety is of the utmost significance. In this study, a deep learning-based technique is presented for identifying safety gear worn by construction workers, such as helmets, goggles, jackets, gloves, and footwear. The suggested method precisely locates these safety items by using the YOLO v7 (You Only Look Once) object detection algorithm. The dataset utilized in this work consists of labeled images split into training, testing and validation sets. Each image has bounding box labels that indicate where the safety equipment is located within the image. The model is trained to identify and categorize the safety equipment based on the labeled dataset through an iterative training approach. We used custom dataset to train this model. Our trained model performed admirably well, with good precision, recall, and F1-score for safety equipment recognition. Also, the model's evaluation produced encouraging results, with a [email protected] score of 87.7%. The model performs effectively, making it possible to quickly identify safety equipment violations on building sites. A thorough evaluation of the outcomes reveals the model's advantages and points up potential areas for development. By offering an automatic and trustworthy method for safety equipment detection, this research contributes to the fields of computer vision and workplace safety. The proposed deep learning-based approach will increase safety compliance and reduce the risk of accidents in the construction industry.

Keywords: deep learning, safety equipment detection, YOLOv7, computer vision, workplace safety

Procedia PDF Downloads 54

Authors: Amrita Mishra, Tushar Kanti Das

Abstract:

Agriculture is the backbone of the economy of Odisha. Sustainability of agriculture holds the key for the development of Odisha. The Sustainable Development Goals are a framework of 17 goals and 169 targets across social, economical and environmental areas of sustainable development. Among all the seventeen goals the second goal is focusing on the promotion of Sustainable Agriculture. In this research our main aim is also to contribute an understanding of effectiveness of sustainable agriculture as a tool for rural development in the selected tribal district (i.e. Koraput and Malkangiri) of Odisha. These two districts are comes under KBK districts of Odisha which are identified as most backward districts of Odisha. The objectives of our study are to investigate the effect of sustainable agriculture on the lives of tribal farmers, to study whether the farmers are empowered by their participation in sustainable agriculture initiatives to move towards their own vision of development and to study the investment and profit ratio in sustainable agriculture. This research will help in filling the major gaps in sociological studies of sustainable agriculture. This information will helpful for farmers, development organisations, donors and policy makers in formulating the development of effective initiatives and policies to support the development of sustainable agriculture. In this study, we have taken 210 respondents and used various statistical techniques like chi-square test, one-way ANOVA and percentage analysis. This research shows that sustainable agriculture is an effective development strategy that benefits the tribal farmers to move towards their own vision of Good Fortune. The poor farmers who struggle to feed their families and maintain viable livelihoods on shrinking land for them sustainable agriculture are really benefited. The farmers are using homemade pesticides, manure and also getting the seeds from different development organisations and Government. So the investment in Sustainable Agriculture is very less. All farmers said their lives are now better than before. The creation of farmers groups for training and marketing for the produces was shown to be very important for empowerment.

Keywords: sustainable, agriculture, tribal farmers, development, empowerment

Procedia PDF Downloads 155

Authors: Manjitsingh Rajput

Abstract:

Visual communication plays an important role in artificial intelligence (AI) because it enables machines to understand and interpret visual information, similar to how humans do. This abstract explores the importance of visual communication in AI and emphasizes the importance of various applications such as computer vision, object emphasis recognition, image classification and autonomous systems. In going deeper, with deep learning techniques and neural networks that modify visual understanding, In addition to AI programming, the abstract discusses challenges facing visual interfaces for AI, such as data scarcity, domain optimization, and interpretability. Visual communication and other approaches, such as natural language processing and speech recognition, have also been explored. Overall, this abstract highlights the critical role that visual communication plays in advancing AI capabilities and enabling machines to perceive and understand the world around them. The abstract also explores the integration of visual communication with other modalities like natural language processing and speech recognition, emphasizing the critical role of visual communication in AI capabilities. This methodology explores the importance of visual communication in AI development and implementation, highlighting its potential to enhance the effectiveness and accessibility of AI systems. It provides a comprehensive approach to integrating visual elements into AI systems, making them more user-friendly and efficient. In conclusion, Visual communication is crucial in AI systems for object recognition, facial analysis, and augmented reality, but challenges like data quality, interpretability, and ethics must be addressed. Visual communication enhances user experience, decision-making, accessibility, and collaboration. Developers can integrate visual elements for efficient and accessible AI systems.

Keywords: visual communication AI, computer vision, visual aid in communication, essence of visual communication.

Procedia PDF Downloads 69

Authors: Abdelsalam Gomaa, Mahjabeen Ramzan, Tooba Ali Akbar, Huma Nadeem

Abstract:

The use of tobacco and the health risks linked to it are well known in this day and age due to the presence of easily available information through the internet. The media is a powerful platform that is used by many anti-smoking awareness campaigns to reach their target audience; yet, it has been found that adolescents are taking up smoking every passing day. Half of this smoking population of youngsters resides in Asia alone, which includes Qatar, the focus country of this study. As smoking happens to be one of the largest avoidable causes of serious diseases like cancers and heart problems, children are taking up smoking at an alarming rate everywhere including Qatar. Importance of the health of the citizens of Qatar is one of the pillars of the Qatar vision 2030, which is to ensure a healthy population, both physically and mentally. Since the youth makes up a significant percentage of the population and in order to achieve the health objectives of the Qatar vision 2030, it is essential to ensure the health and well-being of this part of the population of the country as they are the future of Qatar. Children, especially boys who tend to be more aggressive by nature, are highly likely to develop behavioral and health issues due to smoking at an early age. Research conducted around the world has also emphasized on this association between the smokers developing a bad behaviour as well as poor social communication skills. However, due to lack of research into this association, very little is known about the extent to which smoking impacts the children’s academics, health and behaviour. Moreover, a study of this nature has not yet been conducted in Qatar previously as most of the studies focus on adult smokers and ways to minimize the number of smoking habits in universities and workplaces. This study solely focuses on identifying a relationship between smoking and its impacts on the adolescents by conducting a research on different schools across Qatar.

Keywords: adolescents, modelling techniques, Qatar, smoking

Procedia PDF Downloads 221

Authors: Shao-Di Feng, Von-Wun Soo

Abstract:

To generate symbolic music with specific emotions is a challenging task due to symbolic music datasets that have emotion labels are scarce and incomplete. This research aims to generate more refined emotions based on the training datasets that are only labeled with four quadrants in Russel’s 2D emotion model. We focus on the theory of Music Fadernet and map arousal and valence to the low-level attributes, and build a symbolic music generation model by combining transformer and GM-VAE. We adopt an in-attention mechanism for the model and improve it by allowing modulation by conditional information. And we show the music generation model could control the generation of music according to the emotions specified by users in terms of high-level linguistic expression and by manipulating their corresponding low-level musical attributes. Finally, we evaluate the model performance using a pre-trained emotion classifier against a pop piano midi dataset called EMOPIA, and by subjective listening evaluation, we demonstrate that the model could generate music with more refined emotions correctly.

Keywords: music generation, music emotion controlling, deep learning, semi-supervised learning

Procedia PDF Downloads 68

Authors: Prodipta Bhowmik, Poulami Saha, Preety Mehra, Yogesh Soni, Triloki Nath Jha

Abstract:

In India, the smart city concept is growing day by day. So, for smart city development, a better traffic management and monitoring system is a very important requirement. Nowadays, road accidents increase due to more vehicles on the road. Reckless driving is mainly responsible for a huge number of accidents. So, an efficient traffic management system is required for all kinds of roads to control the traffic speed. The speed limit varies from road to road basis. Previously, there was a radar system but due to high cost and less precision, the radar system is unable to become favorable in a traffic management system. Traffic management system faces different types of problems every day and it has become a researchable topic on how to solve this problem. This paper proposed a computer vision and machine learning-based automated system for multiple vehicle detection, tracking, and speed estimation of vehicles using image processing. Detection of vehicles and estimating their speed from a real-time video is tough work to do. The objective of this paper is to detect vehicles and estimate their speed as accurately as possible. So for this, a real-time video is first captured, then the frames are extracted from that video, then from that frames, the vehicles are detected, and thereafter, the tracking of vehicles starts, and finally, the speed of the moving vehicles is estimated. The goal of this method is to develop a cost-friendly system that can able to detect multiple types of vehicles at the same time.

Keywords: OpenCV, Haar Cascade classifier, DLIB, YOLOV3, centroid tracker, vehicle detection, vehicle tracking, vehicle speed estimation, computer vision

Procedia PDF Downloads 63

Authors: Rachelle McFarlane

Abstract:

The Second Decade of Action Road Safety has begun with increased focus on countries who are disproportionately affected by road fatalities. Researchers highlight the low effectiveness of road safety campaigns in Latin America and the Caribbean (LAC) still reporting approximately 130,000 deaths and six million injuries annually. The regional fatality rate 19.2 per 100,000 with heightened concern for persons 15 to 44 years. In 2021, 483 Jamaicans died in 435 crashes, with 33% of these fatalities occurring during Covid-19 curfew hours. The study objective is to conduct a systemic safety review of Jamaican road crashes and provide a framework for its use in complementing traditional methods. The methodology involves the use of the FHWA Systemic Safety Project Selection Tool for analysis. This tool reviews systemwide data in order to identify risk factors across the network associated with severe and fatal crashes, rather that only hotspots. A total of 10,379 crashes with 745 fatalities and serious injuries were reviewed. Of the focus crash types listed, 50% of ‘Pedestrian Accidents’ resulted in fatalities and serious injuries, followed by 32% ‘Bicycle’, 24% ‘Single’ and 12% of ‘Head-on’. This study seeks to understand the associated risk factors with these priority crash types across the network and recommend cost-effective countermeasures across common sites. As we press towards Vision Zero, the inclusion of the systemic safety review method, complementing traditional methods, may create a wider impact in reducing road fatalities and serious injury by targeting issues across network with similarities; focus crash types and contributing factors.

Keywords: systemic safety review, risk factors, road crashes, crash types

Procedia PDF Downloads 74

Authors: Ingy Ibrahim El-Darwish, Neveen Youssef Azmy

Abstract:

The endowment of a healthy environment can be implemented by endorsing sustainable fundamentals. Design of sustainable buildings through recycling of waste, can reduce health problems, provide good environments and contribute to the aesthetically pleasing entourage. Such environments can help in providing energy-saving alternatives to consolidate the principles of sustainability. The poor community awareness and the absence of laws and legislation in Egypt for waste management specifically in residential areas have led to an inability to provide an integrated system for waste management in urban and rural areas. Many problems and environmental challenges face the Egyptian urban environments. From these problems, is the lack of a cohesive vision for waste collection and recycling for energy-saving. The second problem is the lack public awareness of the short term and long term vision of waste management. Bad practices have adversely affected the efficiency of environmental management systems due to lack of urban legislations that codify collection and recycling of residential communities in Egyptian urban environments. Hence, this research tries to address residents on waste management matters to facilitate legislative process on waste collection and classification within residential units and outside them in a preparation phase for recycling in the Egyptian urban environments. In order to achieve this goal, one of the Egyptian communities has been addressed, analyzed and studied. Waste collection, classification, separation and access to recycling places in the urban city are proposed in preparation for a legislation ruling and regulating the process. Hence, sustainable principles are to be achieved.

Keywords: recycling, residential buildings, sustainability, waste

Procedia PDF Downloads 303

Authors: Oluwatosin S. Adeoye

Abstract:

The power quality issues in the world today are critical to the development of different nations. Prosperity of each nation depends on availability of constant power supply. Constant power supply is a major challenge in Africa particularly in Nigeria where the generated power is than thirty percent of the required power. The metrics of power quality are voltage dip, flickers, spikes, harmonics and interruptions. The level of interruptions in Ekiti State was examined through the investigation of the causes of power interruptions in the State. The method used was the collection of data from the Distribution Company, assessment through simple programming as a command for plotting the graphs through the use of MATLAB 2015 depicting the behavioural pattern of the interruption for a period of six months in 2016. The result shows that the interrelationship between the interruptions and development. Recommendations were suggested with the objective of solving the problems being set up by interruptions in the State and these include installation of reactors, automatic voltage regulators and effective tap changing system on the lines, busses and transformer substation respectively.

Keywords: development, frequency, interruption, power, quality

Procedia PDF Downloads 148

Authors: Reetika Sharma, Lalit Tejwani, Himanshu Shekhar, Arun Singhvi

Abstract:

Purpose: Advanced Glaucoma with Failed trab is not an uncommon sight in glaucoma clinic, and such cases usually tend to present with high intraocular pressure (IOP) and advanced cupping, or even glaucomatous atrophy stage. Re-surgery is needed for such cases, and wound modulation poses a major challenge in these cases. We share our experience in this case series with the use of Ologen (collagen matrix implant) along with MMC 0.04% used in surgery. The purpose of the study was to evaluate the efficacy and outcome of collagen matrix implant in re-trabeculectomy in advanced glaucoma cases. Methodology: Eleven eyes of 11 patients (one eye of one patient) underwent re-trabeculectomy surgery with MMC and Ologen. Ologen implant was used in sub scleral and subconjunctival space, as a spacer and wound modulator. In five cases, triple modulation with implant soaked in anti-VEGF was used. Results: All patients had cupping more than 0.9, and one case was GOA. All cases were on maximal medication at presentation and majority were on systemic anti-glaucoma therapy also. Post-surgery, follow-up ranged from 13 – 34 months, and all cases had a follow longer than the gap between previous surgery (which was failed) and re-trab. One case needed AC reformation and one needling was done. Phaco was done at same sitting in four cases. All cases had their IOP lowered post surgery, and vision was maintained in all, however one case was considered as failed re-surgery case. Topical medication was needed in seven cases post-surgery also. Conclusion: Ologen as adjuvant should be considered in all re-trab cases and all high risk and advanced cases, and triple modulation can be next step in these cases. Aggressive IOP control and non- reluctance to continue topical medications post second surgery should be considered in such cases, to give them best possible vision.

Keywords: failed trabeculectomy, ologen, trabeculectomy, advanced glaucoma

Procedia PDF Downloads 318