Search results for: fault detection and classification

Authors: Samiah Alammari, Nassim Ammour

Abstract:

When providing a massive number of tasks successively to a deep learning process, a good performance of the model requires preserving the previous tasks data to retrain the model for each upcoming classification. Otherwise, the model performs poorly due to the catastrophic forgetting phenomenon. To overcome this shortcoming, we developed a successful continual learning deep model for remote sensing hyperspectral image regions classification. The proposed neural network architecture encapsulates two trainable subnetworks. The first module adapts its weights by minimizing the discrimination error between the land-cover classes during the new task learning, and the second module tries to learn how to replicate the data of the previous tasks by discovering the latent data structure of the new task dataset. We conduct experiments on HSI dataset Indian Pines. The results confirm the capability of the proposed method.

Keywords: continual learning, data reconstruction, remote sensing, hyperspectral image segmentation

Procedia PDF Downloads 230

Authors: Ghazi Bouaziz, Abderrahim Derouiche, Damien Brulin, Hélène Pigot, Eric Campo

Abstract:

Automatic ADL classification is a crucial part of ambient assisted living technologies. It allows to monitor the daily life of the elderly and to detect any changes in their behavior that could be related to health problem. But detection of ADLs is a challenge, especially because each person has his/her own rhythm for performing them. Therefore, we used a correlation matrix to extract custom rules that enable to detect ADLs, including eating activity. Data collected from 3 different individuals between 35 and 105 days allows the extraction of personalized eating patterns. The comparison of the results of the process of eating activity extracted from the correlation matrices with the declarative data collected during the survey shows an accuracy of 90%.

Keywords: elderly monitoring, ADL identification, matrix correlation, meal-taking activity

Procedia PDF Downloads 76

Authors: Weitao Lin

Abstract:

To quickly extract adequate information from large-scale unstructured text data, this paper studies the representation of events in Chinese scenarios and performs the regularized abstraction. It proposes a Chinese event detection technique based on dependency parsing and rule matching. The method first performs dependency parsing on the original utterance, then performs pattern matching at the word or phrase granularity based on the results of dependent syntactic analysis, filters out the utterances with prominent non-event characteristics, and obtains the final results. The experimental results show the effectiveness of the method.

Keywords: natural language processing, Chinese event detection, rules matching, dependency parsing

Procedia PDF Downloads 118

Authors: Neda Orak, Mostafa Zarei

Abstract:

Environmental assessment is an important session in environment management. Since various methods and techniques have been produces and implemented. Remote sensing (RS) is widely used in many scientific and research fields such as geology, cartography, geography, agriculture, forestry, land use planning, environment, etc. It can show earth surface objects cyclical changes. Also, it can show earth phenomena limits on basis of electromagnetic reflectance changes and deviations records. The research has been done on mangrove forests assessment by RS techniques. Mangrove forests quantitative analysis in Basatin and Bidkhoon estuaries was the aim of this research. It has been done by Landsat satellite images from 1975- 2013 and match to ground control points. This part of mangroves are the last distribution in northern hemisphere. It can provide a good background to improve better management on this important ecosystem. Landsat has provided valuable images to earth changes detection to researchers. This research has used MSS, TM, +ETM, OLI sensors from 1975, 1990, 2000, 2003-2013. Changes had been studied after essential corrections such as fix errors, bands combination, georeferencing on 2012 images as basic image, by maximum likelihood and IPVI Index. It was done by supervised classification. 2004 google earth image and ground points by GPS (2010-2012) was used to compare satellite images obtained changes. Results showed mangrove area in bidkhoon was 1119072 m2 by GPS and 1231200 m2 by maximum likelihood supervised classification and 1317600 m2 by IPVI in 2012. Basatin areas is respectively: 466644 m2, 88200 m2, 63000 m2. Final results show forests have been declined naturally. It is due to human activities in Basatin. The defect was offset by planting in many years. Although the trend has been declining in recent years again. So, it mentioned satellite images have high ability to estimation all environmental processes. This research showed high correlation between images and indexes such as IPVI and NDVI with ground control points.

Keywords: IPVI index, Landsat sensor, maximum likelihood supervised classification, Nayband National Park

Procedia PDF Downloads 272

Authors: Xiaoyan Dai, Hsieh Yisan

Abstract:

Barcode processing of captured images is a huge challenge, as different shooting conditions can result in different barcode appearances. This paper proposes a deep learning-based barcode detection using synthetic-to-real data augmentation. We first augment barcodes themselves; we then augment images containing the barcodes to generate a large variety of data that is close to the actual shooting environments. Comparisons with previous works and evaluations with our original data show that this approach achieves state-of-the-art performance in various real images. In addition, the system uses hybrid resolution for barcode “scan” and is applicable to real-time applications.

Keywords: barcode detection, data augmentation, deep learning, image-based processing

Procedia PDF Downloads 140

Authors: Jularat Chumnaul

Abstract:

In forensic science, corpses from various homicides are different; there are both complete and incomplete, depending on causes of death or forms of homicide. For example, some corpses are cut into pieces, some are camouflaged by dumping into the river, some are buried, some are burned to destroy the evidence, and others. If the corpses are incomplete, it can lead to the difficulty of personally identifying because some tissues and bones are destroyed. To specify gender of the corpses from skeletal remains, the most precise method is DNA identification. However, this method is costly and takes longer so that other identification techniques are used instead. The first technique that is widely used is considering the features of bones. In general, an evidence from the corpses such as some pieces of bones, especially the skull and pelvis can be used to identify their gender. To use this technique, forensic scientists are required observation skills in order to classify the difference between male and female bones. Although this technique is uncomplicated, saving time and cost, and the forensic scientists can fairly accurately determine gender by using this technique (apparently an accuracy rate of 90% or more), the crucial disadvantage is there are only some positions of skeleton that can be used to specify gender such as supraorbital ridge, nuchal crest, temporal lobe, mandible, and chin. Therefore, the skeletal remains that will be used have to be complete. The other technique that is widely used for gender specifying in forensic science and archeology is skeletal measurements. The advantage of this method is it can be used in several positions in one piece of bones, and it can be used even if the bones are not complete. In this study, the classification and cluster analysis are applied to this technique, including the Kth Nearest Neighbor Classification, Classification Tree, Ward Linkage Cluster, K-mean Cluster, and Two Step Cluster. The data contains 507 particular individuals and 9 skeletal measurements (diameter measurements), and the performance of five methods are investigated by considering the apparent error rate (APER). The results from this study indicate that the Two Step Cluster and Kth Nearest Neighbor method seem to be suitable to specify gender from human skeletal remains because both yield small apparent error rate of 0.20% and 4.14%, respectively. On the other hand, the Classification Tree, Ward Linkage Cluster, and K-mean Cluster method are not appropriate since they yield large apparent error rate of 10.65%, 10.65%, and 16.37%, respectively. However, there are other ways to evaluate the performance of classification such as an estimate of the error rate using the holdout procedure or misclassification costs, and the difference methods can make the different conclusions.

Keywords: skeletal measurements, classification, cluster, apparent error rate

Procedia PDF Downloads 235

Authors: Hoshang Kolivand, Mahyar Kolivand, Mohd Shahrizal Sunar, Mohd Azhar M. Arsad

Abstract:

Real-time shadow generation in virtual environments and Augmented Reality (AR) was always a hot topic in the last three decades. Lots of calculation for shadow generation among AR needs a fast algorithm to overcome this issue and to be capable of implementing in any real-time rendering. In this paper, a silhouette detection algorithm is presented to generate shadows for AR systems. Δ+ algorithm is presented based on extending edges of occluders to recognize which edges are silhouettes in the case of real-time rendering. An accurate comparison between the proposed algorithm and current algorithms in silhouette detection is done to show the reduction calculation by presented algorithm. The algorithm is tested in both virtual environments and AR systems. We think that this algorithm has the potential to be a fundamental algorithm for shadow generation in all complex environments.

Keywords: silhouette detection, shadow volumes, real-time shadows, rendering, augmented reality

Procedia PDF Downloads 423

Authors: Jarek Krajewski, David Daxberger

Abstract:

We propose a stress detection method based on an RGB camera using heart rate detection, also known as Photoplethysmography Imaging (PPGI). This technique focuses on the measurement of the small changes in skin colour caused by blood perfusion. A stationary lab setting with simulated video conferences is chosen using constant light conditions and a sampling rate of 30 fps. The ground truth measurement of heart rate is conducted with a common PPG system. The proposed approach for pulse peak detection is based on a machine learning-based approach, applying brute force feature extraction for the prediction of heart rate pulses. The statistical analysis showed good agreement (correlation r = .79, p<0.05) between the reference heart rate system and the proposed method. Based on these findings, the proposed method could provide a reliable, low-cost, and contactless way of measuring HR parameters in daily-life environments.

Keywords: heart rate, PPGI, machine learning, brute force feature extraction

Procedia PDF Downloads 113

Authors: Javier Alfonso Bonilla-Chávez, Angélica Lozano

Abstract:

In North America, more than 2,000 people annually die in accidents related to railroad tracks. In 2020, collisions at grade crossings were the main cause of deaths related to railway accidents in Mexico. Railway networks have constant interaction with motor transport users, cyclists, and pedestrians, mainly in grade crossings, where is the greatest vulnerability and risk of accidents. Usually, accidents at grade crossings are directly related to risky behavior and non-compliance with regulations by motorists, cyclists, and pedestrians, especially in developing countries. Around the world, countries classify these crossings in different ways. In Mexico, according to their dangerousness (high, medium, or low), types A, B and C have been established, recommending for each one different type of auditive and visual signaling and gates, as well as horizontal and vertical signaling. This classification is based in a weighting, but regrettably, it is not explained how the weight values were obtained. A review of the variables and the current approach for the grade crossing classification is required, since it is inadequate for some crossings. In contrast, North America (USA and Canada) and European countries consider a broader classification so that attention to each crossing is addressed more precisely and equipment costs are adjusted. Lack of a proper classification, could lead to cost overruns in the equipment and a deficient operation. To exemplify the lack of a good classification, six crossings are studied, three located in the rural area of Mexico and three in Mexico City. These cases show the need of: improving the current regulations, improving the existing infrastructure, and implementing technological systems, including informative signals with nomenclature of the involved crossing and direct telephone line for reporting emergencies. This implementation is unaffordable for most municipal governments. Also, an inventory of the most dangerous grade crossings in urban and rural areas must be obtained. Then, an approach for improving the classification of grade crossings is suggested. This approach must be based on criteria design, characteristics of adjacent roads or intersections which can influence traffic flow through the crossing, accidents related to motorized and non-motorized vehicles, land use and land management, type of area, and services and economic activities in the zone where the grade crossings is located. An expanded classification of grade crossing in Mexico could reduce accidents and improve the efficiency of the railroad.

Keywords: accidents, grade crossing, railroad, traffic safety

Procedia PDF Downloads 93

Authors: R. Biondi, G. Dys, G. Ferone, T. Renard, M. Zysman

Abstract:

This paper describes an automated implementable system for impulsive signals detection and recognition. The system uses a Digital Signal Processing device for the detection and identification process. Here the system analyses the signals in real time in order to produce a particular response if needed. The system analyses the signals in real time in order to produce a specific output if needed. Detection is achieved through normalizing the inputs and comparing the read signals to a dynamic threshold and thus avoiding detections linked to loud or fluctuating environing noise. Identification is done through neuronal network algorithms. As a setup our system can receive signals to “learn” certain patterns. Through “learning” the system can recognize signals faster, inducing flexibility to new patterns similar to those known. Sound is captured through a simple jack input, and could be changed for an enhanced recording surface such as a wide-area recorder. Furthermore a communication module can be added to the apparatus to send alerts to another interface if needed.

Keywords: sound detection, impulsive signal, background noise, neural network

Procedia PDF Downloads 300

Authors: P. S. Jagadeesh Kumar, Yang Yung, Wenli Hu

Abstract:

Speech emotion classification is a dominant research field in finding a sturdy and profligate classifier appropriate for different real-life applications. This effort accentuates on classifying different emotions from speech signal quarried from the features related to pitch, formants, energy contours, jitter, shimmer, spectral, perceptual and temporal features. Tensor deep stacking neural networks were supported to examine the factors that influence the classification success rate. Facial electromyography signals were composed of several forms of focuses in a controlled atmosphere by means of audio-visual stimuli. Proficient facial electromyography signals were pre-processed using moving average filter, and a set of arithmetical features were excavated. Extracted features were mapped into consistent emotions using bilinear mapping. With facial electromyography signals, a database comprising diverse emotions will be exposed with a suitable fine-tuning of features and training data. A success rate of 92% can be attained deprived of increasing the system connivance and the computation time for sorting diverse emotional states.

Keywords: speech emotion classification, tensor deep stacking neural networks, facial electromyography, bilinear mapping, audio-visual stimuli

Procedia PDF Downloads 229

Authors: K. Padmavathi, K. Sri Ramakrishna

Abstract:

This study describes a procedure for the detection of Left and Right Bundle Branch Block (LBBB and RBBB) ECG patterns using spectral Coherence(SC) technique and LM Neural Network. The Coherence function finds common frequencies between two signals and evaluate the similarity of the two signals. The QT variations of Bundle Blocks are observed in lead V1 of ECG. Spectral Coherence technique uses Welch method for calculating PSD. For the detection of normal and Bundle block beats, SC output values are given as the input features for the LMNN classifier. Overall accuracy of LMNN classifier is 99.5 percent. The data was collected from MIT-BIH Arrhythmia database.

Keywords: bundle block, SC, LMNN classifier, welch method, PSD, MIT-BIH, arrhythmia database

Procedia PDF Downloads 257

Authors: Li Kewen, Su Zhaoxin, Wang Xingmou, Zhu Jian Bing

Abstract:

Conventional reservoir prediction methods ar not sufficient to explore the implicit relation between seismic attributes, and thus data utilization is low. In order to improve the predictive classification accuracy of reservoir lithology, this paper proposes a deep learning lithology prediction method based on ResNet (Residual Neural Network) and SENet (Squeeze-and-Excitation Neural Network). The neural network model is built and trained by using seismic attribute data and lithology data of Shengli oilfield, and the nonlinear mapping relationship between seismic attribute and lithology marker is established. The experimental results show that this method can significantly improve the classification effect of reservoir lithology, and the classification accuracy is close to 70%. This study can effectively predict the lithology of undrilled area and provide support for exploration and development.

Keywords: convolutional neural network, lithology, prediction of reservoir, seismic attributes

Procedia PDF Downloads 159

Authors: Regina Chua

Abstract:

To reduce the costs of re-fielding a large survey, a Random Forest classifier was applied to measure the accuracy of classifying individuals into their assigned segments with the fewest possible questions. Given a long survey, one needed to determine the most predictive ten or fewer questions that would accurately assign new individuals to custom segments. Furthermore, the solution needed to be quick in its classification and usable in non-Python environments. In this paper, a supervised Random Forest classifier was modeled on a dataset with 7,000 individuals, 60 questions, and 254 features. The Random Forest consisted of an iterative collection of individual decision trees that result in a predicted segment with robust precision and recall scores compared to a single tree. A random 70-30 stratified sampling for training the algorithm was used, and accuracy trade-offs at different depths for each segment were identified. Ultimately, the Random Forest classifier performed at 87% accuracy at a depth of 10 with 20 instead of 254 features and 10 instead of 60 questions. With an acceptable accuracy in prioritizing feature selection, new tools were developed for non-Python environments: a worksheet with a formulaic version of the algorithm and an embedded function to predict the segment of an individual in real-time. Random Forest was determined to be an optimal classification model by its feature selection, performance, processing speed, and flexible application in other environments.

Keywords: machine learning, supervised learning, data science, random forest, classification, prediction, predictive modeling

Procedia PDF Downloads 79

Authors: José A. Menezes, Giordano Cabral, Bruno T. Gomes

Abstract:

Choosing good features is an essential part of machine learning. Recent techniques aim to automate this process. For instance, feature learning intends to learn the transformation of raw data into a useful representation to machine learning tasks. In automatic audio classification tasks, this is interesting since the audio, usually complex information, needs to be transformed into a computationally convenient input to process. Another technique tries to generate features by searching a feature space. Genetic algorithms, for instance, have being used to generate audio features by combining or modifying them. We find this approach particularly interesting and, despite the undeniable advances of feature learning approaches, we wanted to take a step forward in the use of genetic algorithms to find audio features, combining them with more conventional methods, like PCA, and inserting search control mechanisms, such as constraints over a confusion matrix. This work presents the results obtained on particular audio classification problems.

Keywords: feature generation, feature learning, genetic algorithm, music information retrieval

Procedia PDF Downloads 415

Authors: AlHanoof A. Alharbi, Fatima M. Alamoudi, Razan A. Albrahim, Sarah F. Alharbi, Abdullah M Almuhaideb, Norah A. Almubairik, Abdulrahman Alharby, Naya M. Nagy

Abstract:

Recently, drones received a rapid interest in different industries worldwide due to its powerful impact. However, limitations still exist in this emerging technology, especially privacy violation. These aircrafts consistently threaten the security of entities by entering restricted areas accidentally or deliberately. Therefore, this research project aims to develop drone detection and prevention mechanism to protect the restricted area. Until now, none of the solutions have met the optimal requirements of detection which are cost-effectiveness, high accuracy, long range, convenience, unaffected by noise and generalization. In terms of prevention, the existing methods are focusing on impractical solutions such as catching a drone by a larger drone, training an eagle or a gun. In addition, the practical solutions have limitations, such as the No-Fly Zone and PITBULL jammers. According to our study and analysis of previous related works, none of the solutions includes detection and prevention at the same time. The proposed solution is a combination of detection and prevention methods. To implement the detection system, a passive radar will be used to properly identify the drone against any possible flying objects. As for the prevention, jamming signals and forceful safe landing of the drone integrated together to stop the drone’s operation. We believe that applying this mechanism will limit the drone’s invasion of privacy incidents against highly restricted properties. Consequently, it effectively accelerates drones‘ usages at personal and governmental levels.

Keywords: detection, drone, jamming, prevention, privacy, RF, radar, UAV

Procedia PDF Downloads 192

Authors: Olena Kavats, Dmitry Khramov, Kateryna Sergieieva, Vladimir Vasyliev, Iurii Kavats

Abstract:

Geoinformation technologies of space agromonitoring are a means of operative decision making support in the tasks of managing the agricultural sector of the economy. Existing technologies use satellite images in the optical range of electromagnetic spectrum. Time series of optical images often contain gaps due to the presence of clouds and haze. A geoinformation technology is created. It allows to fill gaps in time series of optical images (Sentinel-2, Landsat-8, PROBA-V, MODIS) with radar survey data (Sentinel-1) and use information about agrometeorological conditions of the growing season for individual monitoring years. The technology allows to perform crop classification and mapping for spring-summer (winter and spring crops) and autumn-winter (winter crops) periods of vegetation, monitoring the dynamics of crop state seasonal changes, crop yield forecasting. Crop classification is based on supervised classification algorithms, takes into account the peculiarities of crop growth at different vegetation stages (dates of sowing, emergence, active vegetation, and harvesting) and agriculture land state characteristics (row spacing, seedling density, etc.). A catalog of samples of the main agricultural crops (Ukraine) is created and crop spectral signatures are calculated with the preliminary removal of row spacing, cloud cover, and cloud shadows in order to construct time series of crop growth characteristics. The obtained data is used in grain crop growth tracking and in timely detection of growth trends deviations from reference samples of a given crop for a selected date. Statistical models of crop yield forecast are created in the forms of linear and nonlinear interconnections between crop yield indicators and crop state characteristics (temperature, precipitation, vegetation indices, etc.). Predicted values of grain crop yield are evaluated with an accuracy up to 95%. The developed technology was used for agricultural areas monitoring in a number of Great Britain and Ukraine regions using EOS Crop Monitoring Platform (https://crop-monitoring.eos.com). The obtained results allow to conclude that joint use of Sentinel-1 and Sentinel-2 images improve separation of winter crops (rapeseed, wheat, barley) in the early stages of vegetation (October-December). It allows to separate successfully the soybean, corn, and sunflower sowing areas that are quite similar in their spectral characteristics.

Keywords: geoinformation technology, crop classification, crop yield prediction, agricultural monitoring, EOS Crop Monitoring Platform

Procedia PDF Downloads 426

Authors: Nicholas Milburn, Yu Liang, Dalei Wu

Abstract:

Athlete performance scoring within the climbing do-main presents interesting challenges as the sport does not have an objective way to assign skill. Assessing skill levels within any sport is valuable as it can be used to mark progress while training, and it can help an athlete choose appropriate climbs to attempt. Machine learning-based methods are popular for complex problems like this. The dataset available was composed of dynamic force data recorded during climbing; however, this dataset came with challenges such as data scarcity, imbalance, and it was temporally heterogeneous. Investigated solutions to these challenges include data augmentation, temporal normalization, conversion of time series to the spectral domain, and cross validation strategies. The investigated solutions to the classification problem included light weight machine classifiers KNN and SVM as well as the deep learning with CNN. The best performing model had an 80% accuracy. In conclusion, there seems to be enough information within climbing force data to accurately categorize climbers by skill.

Keywords: classification, climbing, data imbalance, data scarcity, machine learning, time sequence

Procedia PDF Downloads 128

Authors: Thanakrit Taweesoontorn, Sarucha Yanyong, Poom Konghuayrob

Abstract:

Robotic arms have gained popularity in various industries due to their accuracy and efficiency. This research proposes a method for bin-picking tasks using the Cobot, combining the YOLOv5 CNNs model for object detection and pose estimation with traditional feature detection (FAST), feature description (BRISK), and matching algorithms. By integrating these algorithms and utilizing a small-scale depth sensor camera for capturing depth and color images, the system achieves real-time object detection and accurate pose estimation, enabling the robotic arm to pick objects correctly in both position and orientation. Furthermore, the proposed method is implemented within the ROS framework to provide a seamless platform for robotic control and integration. This integration of robotics, cameras, and AI technology contributes to the development of industrial robotics, opening up new possibilities for automating challenging tasks and improving overall operational efficiency.

Keywords: robotic vision, image processing, applications of robotics, artificial intelligent

Procedia PDF Downloads 70

Authors: Yusuf Ulaş Kabukçu, Si̇nan Çeli̇k, Onur Salan, Mai̇de Altuntaş, Mert Can Dalkiran, Gökseni̇n Bozdağ, Metehan Bulut, Fati̇h Yaman

Abstract:

This study presents a more compact implementation of the 2.4GHz MIT Coffee Can Doppler Radar for 2.6GHz operating frequency. The main difference of our prototype depends on the use of microstrip antennas which makes it possible to transport with a small robotic vehicle. We have designed our radar system with two different channels: Tx and Rx. The system mainly consists of Voltage Controlled Oscillator (VCO) source, low noise amplifiers, microstrip antennas, splitter, mixer, low pass filter, and necessary RF connectors with cables. The two microstrip antennas, one is element for transmitter and the other one is array for receiver channel, was designed, fabricated and verified by experiments. The system has two operation modes: speed detection and range detection. If the switch of the operation mode is ‘Off’, only CW signal transmitted for speed measurement. When the switch is ‘On’, CW is frequency-modulated and range detection is possible. In speed detection mode, high frequency (2.6 GHz) is generated by a VCO, and then amplified to reach a reasonable level of transmit power. Before transmitting the amplified signal through a microstrip patch antenna, a splitter used in order to compare the frequencies of transmitted and received signals. Half of amplified signal (LO) is forwarded to a mixer, which helps us to compare the frequencies of transmitted and received (RF) and has the IF output, or in other words information of Doppler frequency. Then, IF output is filtered and amplified to process the signal digitally. Filtered and amplified signal showing Doppler frequency is used as an input of audio input of a computer. After getting this data Doppler frequency is shown as a speed change on a figure via Matlab script. According to experimental field measurements the accuracy of speed measurement is approximately %90. In range detection mode, a chirp signal is used to form a FM chirp. This FM chirp helps to determine the range of the target since only Doppler frequency measured with CW is not enough for range detection. Such a FMCW Doppler radar may be used in border security of the countries since it is capable of both speed and range detection.

Keywords: doppler radar, FMCW, range detection, speed detection

Procedia PDF Downloads 379

Authors: Chang, Ya-Ju, Hsieh, Pei-Hsin, Wu, Jui-Chuang, Chen-Yang, Yui Whei

Abstract:

A mesoporous silica material was prepared to apply to the lateral-flow immunochromatography for detecting a model biosample. The probe antibody is immobilized on the silica surface as the test line to capture its affinity antigen, which laterally flows through the chromatography strips. The antigen is labeled with nano-gold particles, such that the detection can be visually read out from the test line without instrument aids. The result reveals that the mesoporous material provides a vast area for immobilizing the detection probes. Biosening surfaces corresponding with a positive proportion of detection signals is obtained with the biosample loading.

Keywords: mesoporous silica, immunochromatography, lateral-flow strips, biosensors, nano-gold particles

Procedia PDF Downloads 590

Authors: Aleksandra Zysk, Pawel Badura

Abstract:

Recognizing and controlling vocal registers during singing is a difficult task for beginner vocalist. It requires among others identifying which part of natural resonators is being used when a sound propagates through the body. Thus, an application has been designed allowing for sound recording, automatic vocal register recognition (VRR), and a graphical user interface providing real-time visualization of the signal and recognition results. Six spectral features are determined for each time frame and passed to the support vector machine classifier yielding a binary decision on the head or chest register assignment of the segment. The classification training and testing data have been recorded by ten professional female singers (soprano, aged 19-29) performing sounds for both chest and head register. The classification accuracy exceeded 93% in each of various validation schemes. Apart from a hard two-class clustering, the support vector classifier returns also information on the distance between particular feature vector and the discrimination hyperplane in a feature space. Such an information reflects the level of certainty of the vocal register classification in a fuzzy way. Thus, the designed recognition and training application is able to assess and visualize the continuous trend in singing in a user-friendly graphical mode providing an easy way to control the vocal emission.

Keywords: classification, singing, spectral analysis, vocal emission, vocal register

Procedia PDF Downloads 289

Authors: Rao Singh Priyanka, Jayangondaperumal R.

Abstract:

The Himalayan active fault has the potential to generate great earthquakes in the future, posing a biggest existential threat to humans in the Himalayan and adjacent region. Quantitative evaluation of accumulated and released interseismic strain is crucial to assess the magnitude and spatio-temporal variability of future great earthquakes along the Himalayan arc. To mitigate the destruction and hazards associated with such earthquakes, it is important to understand their recurrence cycle. The eastern Himalayan and Indo-Burman plate boundary systems offers an oblique convergence across two orthogonal plate boundaries, resulting in a zone of distributed deformation both within and away from the plate boundary and clockwise rotation of fault-bounded blocks. This seismically active region has poorly documented historical archive of the past large earthquakes. Thus, paleoseismologicalstudies confirm the surface rupture evidences of the great continental earthquakes (Mw ≥ 8) along the Himalayan Frontal Thrust (HFT), which along with the Geodetic studies, collectively provide the crucial information to understand and assess the seismic potential. These investigations reveal the rupture of 3/4th of the HFT during great events since medieval time but with debatable opinions for the timing of events due to unclear evidences, ignorance of transverse segment boundaries, and lack of detail studies. Recent paleoseismological investigations in the eastern Himalaya and Mishmi ranges confirms the primary surface ruptures of the 1950 C.E. great earthquake (M>8). However, a seismic gap exists between the 1714 C.E. and 1950 C.E. Assam earthquakes that did not slip since 1697 C.E. event. Unlike the latest large blind 2015 Gorkha earthquake (Mw 7.8), the 1950 C.E. event is not triggered by a large event of 1947 C.E. that occurred near the western edge of the great upper Assam event. Moreover, the western segment of the eastern Himalayadid not witness any surface breaking earthquake along the HFT for over the past 300 yr. The frontal fault excavations reveal that during the 1950 earthquake, ~3.1-m-high scarp along the HFT was formed due to the co-seismic slip of 5.5 ± 0.7 m at Pasighat in the Eastern Himalaya and a 10-m-high-scarp at a Kamlang Nagar along the Mishmi Thrust in the Eastern Himalayan Syntaxis is an outcome of a dip-slip displacement of 24.6 ± 4.6 m along a 25 ± 5°E dipping fault. This event has ruptured along the two orthogonal fault systems in the form of oblique thrust fault mechanism. Approx. 130 km west of Pasighat site, the Himebasti village has witnessed two earthquakes, the historical 1697 Sadiya earthquake, and the 1950 event, with a cumulative dip-slip displacement of 15.32 ± 4.69 m. At Niglok site, Arunachal Pradesh, a cumulative slip of ~12.82 m during at least three events since pre 19585 B.P. has produced ~6.2-m high scarp while the youngest scarp of ~2.4-m height has been produced during 1697 C.E. The site preserves two deformational events along the eastern HFT, providing an idea of last serial ruptures at an interval of ~850 yearswhile the successive surface rupturing earthquakes lacks in the Mishmi Range to estimate the recurrence cycle.

Keywords: paleoseismology, surface rupture, recurrence interval, Eastern Himalaya

Procedia PDF Downloads 67

Authors: Mobarok Hossain Bhuyain

Abstract:

Video surveillance processing has attracted various security fields transforming it into one of the leading research fields. Today's demand for detection and tracking of human mobility for security is very useful for human security, such as in crowded areas. Accordingly, video surveillance technology has seen a rapid advancement in recent years, with algorithms analyzing the behavior of people under surveillance automatically. The main motivation of this research focuses on the detection and tracking of the elderly and socially vulnerable people in crowded areas. Degenerate people are a major health concern, especially for elderly people and socially vulnerable people. One major disadvantage of video surveillance is the need for continuous monitoring, especially in crowded areas. To assist the security monitoring live surveillance video, image processing, and artificial intelligence methods can be used to automatically send warning signals to the monitoring officers about elderly people and socially vulnerable people.

Keywords: human detection, target tracking, neural network, particle filter

Procedia PDF Downloads 150

Authors: Jonathan Sohn

Abstract:

Deep neural networks (DNNs) have shown state-of-the-art performance for many applications, including computer vision, natural language processing, and speech recognition. Recently, adversarial attacks have been studied in the context of deep neural networks, which aim to alter the results of deep neural networks by modifying the inputs slightly. For example, an adversarial attack on a DNN used for object detection can cause the DNN to miss certain objects. As a result, the reliability of DNNs is undermined by their lack of robustness against adversarial attacks, raising concerns about their use in safety-critical applications such as autonomous driving. In this paper, we focus on studying the adversarial attacks and defenses on DNNs for image classification. There are two types of adversarial attacks studied which are fast gradient sign method (FGSM) attack and projected gradient descent (PGD) attack. A DNN forms decision boundaries that separate the input images into different categories. The adversarial attack slightly alters the image to move over the decision boundary, causing the DNN to misclassify the image. FGSM attack obtains the gradient with respect to the image and updates the image once based on the gradients to cross the decision boundary. PGD attack, instead of taking one big step, repeatedly modifies the input image with multiple small steps. There is also another type of attack called the target attack. This adversarial attack is designed to make the machine classify an image to a class chosen by the attacker. We can defend against adversarial attacks by incorporating adversarial examples in training. Specifically, instead of training the neural network with clean examples, we can explicitly let the neural network learn from the adversarial examples. In our experiments, the digit recognition accuracy on the MNIST dataset drops from 97.81% to 39.50% and 34.01% when the DNN is attacked by FGSM and PGD attacks, respectively. If we utilize FGSM training as a defense method, the classification accuracy greatly improves from 39.50% to 92.31% for FGSM attacks and from 34.01% to 75.63% for PGD attacks. To further improve the classification accuracy under adversarial attacks, we can also use a stronger PGD training method. PGD training improves the accuracy by 2.7% under FGSM attacks and 18.4% under PGD attacks over FGSM training. It is worth mentioning that both FGSM and PGD training do not affect the accuracy of clean images. In summary, we find that PGD attacks can greatly degrade the performance of DNNs, and PGD training is a very effective way to defend against such attacks. PGD attacks and defence are overall significantly more effective than FGSM methods.

Keywords: deep neural network, adversarial attack, adversarial defense, adversarial machine learning

Procedia PDF Downloads 176

Authors: İmren Öztürk Yılmaz, Abdullah Yasin Bilici, Yasin Atalay Candemir

Abstract:

The system consists of 4 main elements: motion system, image acquisition system, image processing software, and control interface. The parts coming out of the production line to enter the image processing system with the conveyor belt at the end of the line. The 3D scanning of the produced part is performed with the laser scanning system integrated into the system entry side. With the 3D scanning method, it is determined at what position and angle the parts enter the system, and according to the data obtained, parameters such as part origin and conveyor speed are calculated with the designed software, and the robot is informed about the position where it will take part. The robot, which receives the information, takes the produced part on the belt conveyor and shows it to high-resolution cameras for quality control. Measurement processes are carried out with a maximum error of 20 microns determined by the experiments.

Keywords: quality control, industry 4.0, image processing, automated fault detection, digital visual inspection

Procedia PDF Downloads 91

Authors: Kamal KC, Zhendong Yin, Dasen Li, Zhilu Wu

Abstract:

Intra-plant disease classification based on leaf images is a challenging computer vision task due to similarities in texture, color, and shape of leaves with a slight variation of leaf spot; and external environmental changes such as lighting and background noises. Deep convolutional neural network (DCNN) has proven to be an effective tool for binary classification. In this paper, two methods for binary classification of diseased plant leaves using DCNN are presented; model created from scratch and transfer learning. Our main contribution is a thorough evaluation of 4 networks created from scratch and transfer learning of 5 pre-trained models. Training and testing of these models were performed on a plant leaf images dataset belonging to 16 distinct classes, containing a total of 22,265 images from 8 different plants, consisting of a pair of healthy and diseased leaves. We introduce a deep CNN model, Optimized MobileNet. This model with depthwise separable CNN as a building block attained an average test accuracy of 99.77%. We also present a fine-tuning method by introducing the concept of a convolutional block, which is a collection of different deep neural layers. Fine-tuned models proved to be efficient in terms of accuracy and computational cost. Fine-tuned MobileNet achieved an average test accuracy of 99.89% on 8 pairs of [healthy, diseased] leaf ImageSet.

Keywords: deep convolution neural network, depthwise separable convolution, fine-grained classification, MobileNet, plant disease, transfer learning

Procedia PDF Downloads 168

Authors: Abdulkader Helwan

Abstract:

Iris cancer, so called intraocular melanoma is a cancer that starts in the iris; the colored part of the eye that surrounds the pupil. There is a need for an accurate and cost-effective iris cancer detection system since the available techniques used currently are still not efficient. The combination of the image processing and artificial neural networks has a great efficiency for the diagnosis and detection of the iris cancer. Image processing techniques improve the diagnosis of the cancer by enhancing the quality of the images, so the physicians diagnose properly. However, neural networks can help in making decision; whether the eye is cancerous or not. This paper aims to develop an intelligent system that stimulates a human visual detection of the intraocular melanoma, so called iris cancer. The suggested system combines both image processing techniques and neural networks. The images are first converted to grayscale, filtered, and then segmented using prewitt edge detection algorithm to detect the iris, sclera circles and the cancer. The principal component analysis is used to reduce the image size and for extracting features. Those features are considered then as inputs for a neural network which is capable of deciding if the eye is cancerous or not, throughout its experience adopted by many training iterations of different normal and abnormal eye images during the training phase. Normal images are obtained from a public database available on the internet, “Mile Research”, while the abnormal ones are obtained from another database which is the “eyecancer”. The experimental results for the proposed system show high accuracy 100% for detecting cancer and making the right decision.

Keywords: iris cancer, intraocular melanoma, cancerous, prewitt edge detection algorithm, sclera

Procedia PDF Downloads 483

Authors: Ali Asheibi

Abstract:

The rapid increase in the size of databases required to store power quality monitoring data has demanded new techniques for analysing and understanding the data. One suggested technique to assist in analysis is data mining. Preparing raw data to be ready for data mining exploration take up most of the effort and time spent in the whole data mining process. Clustering is an important technique in data mining and machine learning in which underlying and meaningful groups of data are discovered. Large amounts of harmonic data have been collected from an actual harmonic monitoring system in a distribution system in Australia for three years. This amount of acquired data makes it difficult to identify operational events that significantly impact the harmonics generated on the system. In this paper, harmonic data preparation processes to better understanding of the data have been presented. Underlying classes in this data has then been identified using clustering technique based on the Minimum Message Length (MML) method. The underlying operational information contained within the clusters can be rapidly visualised by the engineers. The C5.0 algorithm was used for classification and interpretation of the generated clusters.

Keywords: data mining, harmonic data, clustering, classification

Procedia PDF Downloads 230

Authors: Marta D'Onofrio

Abstract:

The debate on the digitalization of the economy has raised questions about how both labor and the regulation of work processes are changing due to the introduction of digital technologies in the productive system. Within the literature, the term ‘digital labor’ is commonly used to identify the impact of digitalization on labor. Despite the wide use of this term, it is still not available an unambiguous definition of it, and this could create confusion in the use of terminology and in the attempts of classification. As a consequence, the purpose of this paper is to provide for a definition and to propose a classification of ‘digital labor’, resorting to the theoretical approach of organizational studies.

Keywords: digital labor, digitalization, data-driven algorithms, big data, organizational studies

Procedia PDF Downloads 135