Search results for: deep learning network

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

Abstract:

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

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

Procedia PDF Downloads 107

Authors: William Whiteley, Jens Gregor

Abstract:

In recent developments of direct neural network based positron emission tomography (PET) reconstruction, two prominent architectures have emerged for converting measurement data into images: 1) networks that contain fully-connected layers; and 2) networks that primarily use a convolutional encoder-decoder architecture. In this paper, we present a multi-modal direct PET reconstruction method called MDPET, which is a hybrid approach that combines the advantages of both types of networks. MDPET processes raw data in the form of sinograms and histo-images in concert with attenuation maps to produce high quality multi-slice PET images (e.g., 8x440x440). MDPET is trained on a large whole-body patient data set and evaluated both quantitatively and qualitatively against target images reconstructed with the standard PET reconstruction benchmark of iterative ordered subsets expectation maximization. The results show that MDPET outperforms the best previously published direct neural network methods in measures of bias, signal-to-noise ratio, mean absolute error, and structural similarity.

Keywords: deep learning, image reconstruction, machine learning, neural network, positron emission tomography

Procedia PDF Downloads 88

Authors: Yuanqing Li, Changyi Lei, Zhaopeng Xue, Zhuo Zheng, Yanbo Long

Abstract:

In agricultural production and breeding, implementing automatic picking robot in orchard farming to reduce human labour and error is challenging. The core function of it is automatic identification based on machine vision. This paper focuses on apple detection and counting in orchards and implements several deep learning methods. Extensive datasets are used and a semi-automatic annotation method is proposed. The proposed deep learning models are in state-of-the-art YOLO family. In view of the essence of the models with various backbones, a multi-dimensional comparison in details is made in terms of counting accuracy, mAP and model memory, laying the foundation for realising automatic precision agriculture.

Keywords: agricultural object detection, deep learning, machine vision, YOLO family

Procedia PDF Downloads 165

Authors: Bin Feng, Zhulin Zong

Abstract:

Radar systems are critical in the modern military, aviation, and maritime operations, and their proper functioning is essential for the success of these operations. However, due to the complexity and sensitivity of radar systems, they are susceptible to various faults that can significantly affect their performance. Traditional radar fault diagnosis strategies rely on expert knowledge and rule-based approaches, which are often limited in effectiveness and require a lot of time and resources. Deep learning has recently emerged as a promising approach for fault diagnosis due to its ability to learn features and patterns from large amounts of data automatically. In this paper, we propose a radar fault diagnosis strategy based on deep learning that can accurately identify and classify faults in radar systems. Our approach uses convolutional neural networks (CNN) to extract features from radar signals and fault classify the features. The proposed strategy is trained and validated on a dataset of measured radar signals with various types of faults. The results show that it achieves high accuracy in fault diagnosis. To further evaluate the effectiveness of the proposed strategy, we compare it with traditional rule-based approaches and other machine learning-based methods, including decision trees, support vector machines (SVMs), and random forests. The results demonstrate that our deep learning-based approach outperforms the traditional approaches in terms of accuracy and efficiency. Finally, we discuss the potential applications and limitations of the proposed strategy, as well as future research directions. Our study highlights the importance and potential of deep learning for radar fault diagnosis. It suggests that it can be a valuable tool for improving the performance and reliability of radar systems. In summary, this paper presents a radar fault diagnosis strategy based on deep learning that achieves high accuracy and efficiency in identifying and classifying faults in radar systems. The proposed strategy has significant potential for practical applications and can pave the way for further research.

Keywords: radar system, fault diagnosis, deep learning, radar fault

Procedia PDF Downloads 52

Authors: Sirilak Areerachakul, Nat Ployong, Supayothin Na Songkla

Abstract:

This research’s objective is to select the model with most accurate value by using Neural Network Technique as a way to filter potential students who enroll in IT course by electronic learning at Suan Suanadha Rajabhat University. It is designed to help students selecting the appropriate courses by themselves. The result showed that the most accurate model was 100 Folds Cross-validation which had 73.58% points of accuracy.

Keywords: artificial neural network, classification, students, e-learning

Procedia PDF Downloads 393

Authors: Chien-kuo Chang, Bo-wei Wu, Yi-yun Tang, Min-chiu Wu

Abstract:

This paper proposes a power equipment diagnosis system based on partial discharge (PD), which is characterized by increasing the readability of experimental data and the convenience of operation. This system integrates a variety of analysis programs of different data formats and different programming languages and then establishes a set of interfaces that can follow and expand the structure, which is also helpful for subsequent maintenance and innovation. This study shows a case of using the developed Convolutional Neural Networks (CNN) to integrate with this system, using the designed model architecture to simplify the complex training process. It is expected that the simplified training process can be used to establish an adaptive deep learning experimental structure. By selecting different test data for repeated training, the accuracy of the identification system can be enhanced. On this platform, the measurement status and partial discharge pattern of each equipment can be checked in real time, and the function of real-time identification can be set, and various training models can be used to carry out real-time partial discharge insulation defect identification and insulation state diagnosis. When the electric power equipment entering the dangerous period, replace equipment early to avoid unexpected electrical accidents.

Keywords: partial discharge, convolutional neural network, partial discharge analysis platform, adaptive deep learning

Procedia PDF Downloads 46

Authors: Sheldon McCall, Miao Yu, Liyun Gong, Shigang Yue, Stefanos Kollias

Abstract:

Falls are a significant health concern for older adults globally, and prompt identification is critical to providing necessary healthcare support. Our study proposes a new fall detection method using computer vision based on modern deep learning techniques. Our approach involves training a trans- former model on a large 2D pose dataset for general action recognition, followed by transfer learning. Specifically, we freeze the first few layers of the trained transformer model and train only the last two layers for fall detection. Our experimental results demonstrate that our proposed method outperforms both classical machine learning and deep learning approaches in fall/non-fall classification. Overall, our study suggests that our proposed methodology could be a valuable tool for identifying falls.

Keywords: healthcare, fall detection, transformer, transfer learning

Procedia PDF Downloads 97

Authors: Vishnu Pratap Singh Kirar, Madhuri Saxena

Abstract:

Advancement in sequence data generation technologies is churning out voluminous omics data and posing a massive challenge to annotate the biological functional features. With so much data available, the use of machine learning methods and tools to make novel inferences has become obvious. Machine learning methods have been successfully applied to a lot of disciplines, including computational biology and bioinformatics. Researchers in computational biology are interested to develop novel machine learning frameworks to classify the huge amounts of biological data. In this proposal, it plan to employ novel machine learning approaches to aid the understanding of how apparently innocuous mutations (in intergenic DNA and at synonymous sites) cause diseases. We are also interested in discovering novel functional sites in the genome and mutations in which can affect a phenotype of interest.

Keywords: genome wide association studies (GWAS), next generation sequencing (NGS), deep learning, omics

Procedia PDF Downloads 63

Authors: Thiago Spilborghs Bueno Meyer, Plinio Thomaz Aquino Junior

Abstract:

Through speech, which privileges the functional and interactive nature of the text, it is possible to ascertain the spatiotemporal circumstances, the conditions of production and reception of the discourse, the explicit purposes such as informing, explaining, convincing, etc. These conditions allow bringing the interaction between humans closer to the human-robot interaction, making it natural and sensitive to information. However, it is not enough to understand what is said; it is necessary to recognize emotions for the desired interaction. The validity of the use of neural networks for feature selection and emotion recognition was verified. For this purpose, it is proposed the use of neural networks and comparison of models, such as recurrent neural networks and deep neural networks, in order to carry out the classification of emotions through speech signals to verify the quality of recognition. It is expected to enable the implementation of robots in a domestic environment, such as the HERA robot from the RoboFEI@Home team, which focuses on autonomous service robots for the domestic environment. Tests were performed using only the Mel-Frequency Cepstral Coefficients, as well as tests with several characteristics of Delta-MFCC, spectral contrast, and the Mel spectrogram. To carry out the training, validation and testing of the neural networks, the eNTERFACE’05 database was used, which has 42 speakers from 14 different nationalities speaking the English language. The data from the chosen database are videos that, for use in neural networks, were converted into audios. It was found as a result, a classification of 51,969% of correct answers when using the deep neural network, when the use of the recurrent neural network was verified, with the classification with accuracy equal to 44.09%. The results are more accurate when only the Mel-Frequency Cepstral Coefficients are used for the classification, using the classifier with the deep neural network, and in only one case, it is possible to observe a greater accuracy by the recurrent neural network, which occurs in the use of various features and setting 73 for batch size and 100 training epochs.

Keywords: emotion recognition, speech, deep learning, human-robot interaction, neural networks

Procedia PDF Downloads 130

Authors: Niousha Bagheri Khulenjani, Mohammad Saniee Abadeh

Abstract:

Learning from very big datasets is a significant problem for most present data mining and machine learning algorithms. MicroRNA (miRNA) is one of the important big genomic and non-coding datasets presenting the genome sequences. In this paper, a hybrid method for the classification of the miRNA data is proposed. Due to the variety of cancers and high number of genes, analyzing the miRNA dataset has been a challenging problem for researchers. The number of features corresponding to the number of samples is high and the data suffer from being imbalanced. The feature selection method has been used to select features having more ability to distinguish classes and eliminating obscures features. Afterward, a Convolutional Neural Network (CNN) classifier for classification of cancer types is utilized, which employs a Genetic Algorithm to highlight optimized hyper-parameters of CNN. In order to make the process of classification by CNN faster, Graphics Processing Unit (GPU) is recommended for calculating the mathematic equation in a parallel way. The proposed method is tested on a real-world dataset with 8,129 patients, 29 different types of tumors, and 1,046 miRNA biomarkers, taken from The Cancer Genome Atlas (TCGA) database.

Keywords: cancer classification, feature selection, deep learning, genetic algorithm

Procedia PDF Downloads 88

Authors: Guantao Bai

Abstract:

Public space is an indispensable and important component of the urban built environment. How to more accurately evaluate the usage characteristics of public space can help improve its spatial quality. Compared to traditional survey methods, computer vision technology based on deep learning has advantages such as dynamic observation and low cost. This study takes the public space of Shaping Park as an example and, based on deep learning computer vision technology, processes and analyzes the image data of the public space to obtain the spatial usage characteristics and spatiotemporal characteristics of the public space. Research has found that the spontaneous activity time in public spaces is relatively random with a relatively short average activity time, while social activities have a relatively stable activity time with a longer average activity time. Computer vision technology based on deep learning can effectively describe the spatial usage characteristics of the research area, making up for the shortcomings of traditional research methods and providing relevant support for creating a good public space.

Keywords: computer vision, deep learning, public spaces, using features

Procedia PDF Downloads 36

Authors: Maulik Parmar, Sumeet Sandhu

Abstract:

The online fashion industry experiences high product return rates. Many returns are because of size/fit mismatches -the size scale on labels can vary across brands, the size parameters may not capture all fit measurements, or the product may have manufacturing defects. Warehouse quality check of garment sizes can be semi-automated to improve speed and accuracy. This paper presents an approach for automatically measuring garment sizes from a single image of the garment -using Deep Learning to learn garment keypoints. The paper focuses on the waist size measurement of jeans and can be easily extended to other garment types and measurements. Experimental results show that this approach can greatly improve the speed and accuracy of today’s manual measurement process.

Keywords: convolutional neural networks, deep learning, distortion, garment measurements, image warping, keypoints

Procedia PDF Downloads 264

Authors: Krishna Mohan Bathula, Fatou Bintou Loucoubar, FNU Kaleemunnisa, Christelle Scharff, Mark Anthony De Castro

Abstract:

Voice recognition algorithms such as automatic speech recognition and text-to-speech systems with African languages can play an important role in bridging the digital divide of Artificial Intelligence in Africa, contributing to the establishment of a fully inclusive information society. This paper proposes a Deep Learning model that can classify the user responses as inputs for an interactive voice response system. A dataset with Wolof language words ‘yes’ and ‘no’ is collected as audio recordings. A two stage Data Augmentation approach is adopted for enhancing the dataset size required by the deep neural network. Data preprocessing and feature engineering with Mel-Frequency Cepstral Coefficients are implemented. Convolutional Neural Networks (CNNs) have proven to be very powerful in image classification and are promising for audio processing when sounds are transformed into spectra. For performing voice response classification, the recordings are transformed into sound frequency feature spectra and then applied image classification methodology using a deep CNN model. The inference model of this trained and reusable Wolof voice response recognition system can be integrated with many applications associated with both web and mobile platforms.

Keywords: automatic speech recognition, interactive voice response, voice response recognition, wolof word classification

Procedia PDF Downloads 85

Authors: Ahmed Al-Adaileh

Abstract:

Renewable energy sources, domestic storage systems, controllable loads and machine learning technologies will be key components of future smart homes management systems. An energy management scheme that uses a Deep Learning (DL) approach to support the smart home management systems, which consist of a standalone photovoltaic system, storage unit, heating ventilation air-conditioning system and a set of conventional and smart appliances, is presented. The objective of the proposed scheme is to apply DL-based machine learning to predict various running parameters within a smart home's environment to achieve maximum comfort levels for occupants, reduced electricity bills, and less dependency on the public grid. The problem is using Reinforcement learning, where decisions are taken based on applying the Continuous-time Markov Decision Process. The main contribution of this research is the proposed framework that applies DL to enhance the system's supervised dataset to offer unlimited chances to effectively support smart home systems. A case study involving a set of conventional and smart appliances with dedicated processing units in an inhabited building can demonstrate the validity of the proposed framework. A visualization graph can show "before" and "after" results.

Keywords: smart homes systems, machine learning, deep learning, Markov Decision Process

Procedia PDF Downloads 159

Authors: Li Hui, Riyadh Hindi

Abstract:

In the United States, bridge construction often employs overhang brackets to support the deck overhang, the weight of fresh concrete, and loads from construction equipment. This approach, however, can lead to significant torsional moments on the exterior girders, potentially causing excessive girder rotation. Such rotations can result in various safety and maintenance issues, including thinning of the deck, reduced concrete cover, and cracking during service. Traditionally, these issues are addressed by installing temporary lateral bracing systems and conducting comprehensive torsional analysis through detailed finite element analysis for the construction of bridge deck overhang. However, this process is often intricate and time-intensive, with the spacing between temporary lateral bracing systems usually relying on the field engineers’ expertise. In this study, a deep neural network model is introduced to limit exterior girder rotation during bridge deck construction. The model predicts the optimal spacing between temporary bracing systems. To train this model, over 10,000 finite element models were generated in SAP2000, incorporating varying parameters such as girder dimensions, span length, and types and spacing of lateral bracing systems. The findings demonstrate that the deep neural network provides an effective and efficient alternative for limiting the exterior girder rotation for bridge deck construction. By reducing dependence on extensive finite element analyses, this approach stands out as a significant advancement in improving safety and maintenance effectiveness in the construction of bridge decks.

Keywords: bridge deck construction, exterior girder rotation, deep learning, finite element analysis

Procedia PDF Downloads 31

Authors: Parisa Golbayani, Dan Wang, Ionut¸ Florescu

Abstract:

In this work we implement machine learning techniques to financial statement reports in order to asses company’s credit rating. Specifically, the work analyzes the performance of four neural network architectures (MLP, CNN, CNN2D, LSTM) in predicting corporate credit rating as issued by Standard and Poor’s. The paper focuses on companies from the energy, financial, and healthcare sectors in the US. The goal of this analysis is to improve application of machine learning algorithms to credit assessment. To accomplish this, the study investigates three questions. First, we investigate if the algorithms perform better when using a selected subset of important features or whether better performance is obtained by allowing the algorithms to select features themselves. Second, we address the temporal aspect inherent in financial data and study whether it is important for the results obtained by a machine learning algorithm. Third, we aim to answer if one of the four particular neural network architectures considered consistently outperforms the others, and if so under which conditions. This work frames the problem as several case studies to answer these questions and analyze the results using ANOVA and multiple comparison testing procedures.

Keywords: convolutional neural network, long short term memory, multilayer perceptron, credit rating

Procedia PDF Downloads 207

Authors: Endrick Barnacin, Jean-Luc Henry, Jimmy Nagau, Jack Molinie

Abstract:

Palynology is a field of interest in many disciplines due to its multiple applications: chronological dating, climatology, allergy treatment, and honey characterization. Unfortunately, the analysis of a pollen slide is a complicated and time consuming task that requires the intervention of experts in the field, which are becoming increasingly rare due to economic and social conditions. That is why the need for automation of this task is urgent. A lot of studies have investigated the subject using different standard image processing descriptors and sometimes hand-crafted ones.In this work, we make a comparative study between classical feature extraction methods (Shape, GLCM, LBP, and others) and Deep Learning (CNN, Autoencoders, Transfer Learning) to perform a recognition task over 80 regional pollen species. It has been found that the use of Transfer Learning seems to be more precise than the other approaches

Keywords: pollens identification, features extraction, pollens classification, automated palynology

Procedia PDF Downloads 102

Authors: A. Kavya Sri, K. Sai Vineela, R. Vanitha, S. Rohith

Abstract:

Large pieces of ice that break from the glaciers are known as icebergs. The threat that icebergs pose to navigation, production of offshore oil and gas services, and underwater pipelines makes their detection crucial. In this project, an automated iceberg tracking method using deep learning techniques and satellite images of icebergs is to be developed. With a temporal resolution of 12 days and a spatial resolution of 20 m, Sentinel-1 (SAR) images can be used to track iceberg drift over the Southern Ocean. In contrast to multispectral images, SAR images are used for analysis in meteorological conditions. This project develops a web-based graphical user interface to detect and track icebergs using sentinel-1 images. To track the movement of the icebergs by using temporal images based on their latitude and longitude values and by comparing the center and area of all detected icebergs. Testing the accuracy is done by precision and recall measures.

Keywords: synthetic aperture radar (SAR), icebergs, deep learning, spatial resolution, temporal resolution

Procedia PDF Downloads 56

Authors: Murad Almadani, Khalil Abu-Hantash, Xinyu Wang, Herbert Jelinek, Kinda Khalaf

Abstract:

Gait analysis is used by healthcare professionals as a tool to gain a better understanding of the movement impairment and track progress. In most circumstances, monitoring patients in their real-life environments with low-cost equipment such as cameras and wearable sensors is more important. Inertial sensors, on the other hand, cannot provide enough information on angular dynamics. This research offers a method for tracking 2D joint coordinates using cutting-edge vision algorithms and a single RGB camera. We provide an end-to-end comprehensive deep learning pipeline for marker-less gait parameter estimation, which, to our knowledge, has never been done before. To make our pipeline function in real-time for real-world applications, we leverage the AlphaPose human posture prediction model and a deep learning transformer. We tested our approach on the well-known GPJATK dataset, which produces promising results.

Keywords: gait analysis, human pose estimation, deep learning, real time gait estimation, AlphaPose, transformer

Procedia PDF Downloads 82

Authors: Gong Zhilin, Jing Yang, Jian Yin

Abstract:

The purpose of this paper is to develop a novel system for credit card fraud detection based on sequential modeling of data using hybrid deep learning models. The projected model encapsulates five major phases are pre-processing, imbalance-data handling, feature extraction, optimal feature selection, and fraud detection with an ensemble classifier. The collected raw data (input) is pre-processed to enhance the quality of the data through alleviation of the missing data, noisy data as well as null values. The pre-processed data are class imbalanced in nature, and therefore they are handled effectively with the K-means clustering-based SMOTE model. From the balanced class data, the most relevant features like improved Principal Component Analysis (PCA), statistical features (mean, median, standard deviation) and higher-order statistical features (skewness and kurtosis). Among the extracted features, the most optimal features are selected with the Self-improved Arithmetic Optimization Algorithm (SI-AOA). This SI-AOA model is the conceptual improvement of the standard Arithmetic Optimization Algorithm. The deep learning models like Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), and optimized Quantum Deep Neural Network (QDNN). The LSTM and CNN are trained with the extracted optimal features. The outcomes from LSTM and CNN will enter as input to optimized QDNN that provides the final detection outcome. Since the QDNN is the ultimate detector, its weight function is fine-tuned with the Self-improved Arithmetic Optimization Algorithm (SI-AOA).

Keywords: credit card, data mining, fraud detection, money transactions

Procedia PDF Downloads 102

Authors: Vidhu Mitha Goutham, Ruma Mukherjee

Abstract:

The low availability of well-trained, unlimited, dynamic-access models for specific languages makes it hard for corporate users to adopt quick translation techniques and incorporate them into product solutions. As translation tasks increasingly require a dynamic sequence learning curve; stable, cost-free opensource models are scarce. We survey and compare current translation techniques and propose a modified sequence to sequence model repurposed with attention techniques. Sequence learning using an encoder-decoder model is now paving the path for higher precision levels in translation. Using a Convolutional Neural Network (CNN) encoder and a Recurrent Neural Network (RNN) decoder background, we use Fairseq tools to produce an end-to-end bilingually trained Spanish-English machine translation model including source language detection. We acquire competitive results using a duo-lingo-corpus trained model to provide for prospective, ready-made plug-in use for compound sentences and document translations. Our model serves a decent system for large, organizational data translation needs. While acknowledging its shortcomings and future scope, it also identifies itself as a well-optimized deep neural network model and solution.

Keywords: attention, encoder-decoder, Fairseq, Seq2Seq, Spanish, translation

Procedia PDF Downloads 148

Authors: Danilo López, Nelson Vera, Luis Pedraza

Abstract:

This paper analyzes fundamental ideas and concepts related to neural networks, which provide the reader a theoretical explanation of Long Short-Term Memory (LSTM) networks operation classified as Deep Learning Systems, and to explicitly present the mathematical development of Backward Pass equations of the LSTM network model. This mathematical modeling associated with software development will provide the necessary tools to develop an intelligent system capable of predicting the behavior of licensed users in wireless cognitive radio networks.

Keywords: neural networks, multilayer perceptron, long short-term memory, recurrent neuronal network, mathematical analysis

Procedia PDF Downloads 383

Authors: Ian Omung'a

Abstract:

Breast cancer remains to be one of the deadliest cancers for women worldwide, with the risk of developing tumors being as high as 50 percent in Sub-Saharan African countries like Kenya. With as many as 42 percent of these cases set to be diagnosed late when cancer has metastasized and or the prognosis has become terminal, Full Field Digital [FFD] Mammography remains an effective screening technique that leads to early detection where in most cases, successful interventions can be made to control or eliminate the tumors altogether. FFD Mammograms have been proven to multiply more effective when used together with Computer-Aided Detection and Diagnosis [CADe] systems, relying on algorithmic implementations of Deep Learning techniques in Computer Vision to carry out deep pattern recognition that is comparable to the level of a human radiologist and decipher whether specific areas of interest in the mammogram scan image portray abnormalities if any and whether these abnormalities are indicative of a benign or malignant tumor. Within this paper, we review emergent Deep Learning techniques that will prove relevant to the development of State-of-The-Art FFD Mammogram CADe systems. These techniques will span self-supervised learning for context-encoded occlusion, self-supervised learning for pre-processing and labeling automation, as well as the creation of a standardized large-scale mammography dataset as a benchmark for CADe systems' evaluation. Finally, comparisons are drawn between existing practices that pre-date these techniques and how the development of CADe systems that incorporate them will be different.

Keywords: breast cancer diagnosis, computer aided detection and diagnosis, deep learning, whole mammogram classfication, ultrasound classification, computer vision

Procedia PDF Downloads 70

Authors: Erick C. Valverde, Wansu Lim

Abstract:

Facial emotion recognition (FER) system has been recently developed for more advanced computer vision applications. The ability to identify human emotions would enable smart healthcare facility to diagnose mental health illnesses (e.g., depression and stress) as well as better human social interactions with smart technologies. The FER system involves two steps: 1) face detection task and 2) facial emotion recognition task. It classifies the human expression in various categories such as angry, disgust, fear, happy, sad, surprise, and neutral. This system requires intensive research to address issues with human diversity, various unique human expressions, and variety of human facial features due to age differences. These issues generally affect the ability of the FER system to detect human emotions with high accuracy. Early stage of FER systems used simple supervised classification task algorithms like K-nearest neighbors (KNN) and artificial neural networks (ANN). These conventional FER systems have issues with low accuracy due to its inefficiency to extract significant features of several human emotions. To increase the accuracy of FER systems, deep learning (DL)-based methods, like convolutional neural networks (CNN), are proposed. These methods can find more complex features in the human face by means of the deeper connections within its architectures. However, the inference speed and computational costs of a DL-based FER system is often disregarded in exchange for higher accuracy results. To cope with this drawback, an optimized DL-based FER system is proposed in this study.An extreme version of Inception V3, known as Xception model, is leveraged by applying different network optimization methods. Specifically, network pruning and quantization are used to enable lower computational costs and reduce memory usage, respectively. To support low resource requirements, a 68-landmark face detector from Dlib is used in the early step of the FER system.Furthermore, a DL compiler is utilized to incorporate advanced optimization techniques to the Xception model to improve the inference speed of the FER system. In comparison to VGG-Net and ResNet50, the proposed optimized DL-based FER system experimentally demonstrates the objectives of the network optimization methods used. As a result, the proposed approach can be used to create an efficient and real-time FER system.

Keywords: deep learning, face detection, facial emotion recognition, network optimization methods

Procedia PDF Downloads 78

Authors: Mirjana Ruppel, Rajendra Persad, Amit Bahl, Sanja Dogramadzi, Chris Melhuish, Lyndon Smith

Abstract:

Multimodal image registration is a profoundly complex task which is why deep learning has been used widely to address it in recent years. However, two main challenges remain: Firstly, the lack of ground truth data calls for an unsupervised learning approach, which leads to the second challenge of defining a feasible loss function that can compare two images of different modalities to judge their level of alignment. To avoid this issue altogether we implement a generative adversarial network consisting of two registration networks GAB, GBA and two discrimination networks DA, DB connected by spatial transformation layers. GAB learns to generate a deformation field which registers an image of the modality B to an image of the modality A. To do that, it uses the feedback of the discriminator DB which is learning to judge the quality of alignment of the registered image B. GBA and DA learn a mapping from modality A to modality B. Additionally, a cycle-consistency loss is implemented. For this, both registration networks are employed twice, therefore resulting in images ˆA, ˆB which were registered to ˜B, ˜A which were registered to the initial image pair A, B. Thus the resulting and initial images of the same modality can be easily compared. A dataset of liver CT and MRI was used to evaluate the quality of our approach and to compare it against learning and non-learning based registration algorithms. Our approach leads to dice scores of up to 0.80 ± 0.01 and is therefore comparable to and slightly more successful than algorithms like SimpleElastix and VoxelMorph.

Keywords: cycle consistency, deformable multimodal image registration, deep learning, GAN

Procedia PDF Downloads 98

Authors: Jie Zhang, Qianyu Guo, Tieyi Zhang, Zhiyong Feng, Xiaohong Li

Abstract:

With the widespread popularity of mobile devices and the development of artificial intelligence (AI), deep learning (DL) has been extensively applied in Android apps. Compared with traditional Android apps (traditional apps), deep learning based Android apps (DL-based apps) need to use more third-party application programming interfaces (APIs) to complete complex DL inference tasks. However, existing methods (e.g., FlowDroid) for detecting sensitive information leakage in Android apps cannot be directly used to detect DL-based apps as they are difficult to detect third-party APIs. To solve this problem, we design DLtrace; a new static information flow analysis tool that can effectively recognize third-party APIs. With our proposed trace and detection algorithms, DLtrace can also efficiently detect privacy leaks caused by sensitive APIs in DL-based apps. Moreover, using DLtrace, we summarize the non-sequential characteristics of DL inference tasks in DL-based apps and the specific functionalities provided by DL models for such apps. We propose two formal definitions to deal with the common polymorphism and anonymous inner-class problems in the Android static analyzer. We conducted an empirical assessment with DLtrace on 208 popular DL-based apps in the wild and found that 26.0% of the apps suffered from sensitive information leakage. Furthermore, DLtrace has a more robust performance than FlowDroid in detecting and identifying third-party APIs. The experimental results demonstrate that DLtrace expands FlowDroid in understanding DL-based apps and detecting security issues therein.

Keywords: mobile computing, deep learning apps, sensitive information, static analysis

Procedia PDF Downloads 125

Authors: Khadija Bouzaachane, El Mahdi El Guarmah

Abstract:

The classification of low-resolution satellite images has been a worthwhile and fertile field that attracts plenty of researchers due to its importance in monitoring geographical areas. It could be used for several purposes such as disaster management, military surveillance, agricultural monitoring. The main objective of this work is to classify efficiently and accurately low-resolution satellite images by using novel technics of deep learning and reinforcement learning. The images include roads, residential areas, industrial areas, rivers, sea lakes, and vegetation. To achieve that goal, we carried out experiments on the sentinel-2 images considering both high accuracy and efficiency classification. Our proposed model achieved a 91% accuracy on the testing dataset besides a good classification for land cover. Focus on the parameter precision; we have obtained 93% for the river, 92% for residential, 97% for residential, 96% for the forest, 87% for annual crop, 84% for herbaceous vegetation, 85% for pasture, 78% highway and 100% for Sea Lake.

Keywords: classification, deep learning, reinforcement learning, satellite imagery

Procedia PDF Downloads 172

Authors: Felipe A. Guth, Shane Ward, Kevin McDonnell

Abstract:

The assessment of disease levels in crop fields is an important and time-consuming task that generally relies on expert knowledge of trained individuals. Image classification in agriculture problems historically has been based on classical machine learning strategies that make use of hand-engineered features in the top of a classification algorithm. This approach tends to not produce results with high accuracy and generalization to the classes classified by the system when the nature of the elements has a significant variability. The advent of deep convolutional neural networks has revolutionized the field of machine learning, especially in computer vision tasks. These networks have great resourcefulness of learning and have been applied successfully to image classification and object detection tasks in the last years. The objective of this work was to propose a new method based on deep learning convolutional neural networks towards the task of disease level monitoring. Common RGB images of winter wheat were obtained during a growing season. Five categories of disease levels presence were produced, in collaboration with agronomists, for the algorithm classification. Disease level tasks performed by experts provided ground truth data for the disease score of the same winter wheat plots were RGB images were acquired. The system had an overall accuracy of 84% on the discrimination of the disease level classes.

Keywords: crop disease assessment, deep learning, precision agriculture, residual neural networks

Procedia PDF Downloads 296

Authors: Alexandre Barbosa de Almeida, Telma Woerle de Lima Soares

Abstract:

Protein structure prediction is a challenging task in the bioinformatics field. The biological function of all proteins majorly relies on the shape of their three-dimensional conformational structure, but less than 1% of all known proteins in the world have their structure solved. This work proposes a deep learning model to address this problem, attempting to predict some aspects of the protein conformations. Throughout a process of multiobjective dominance, a recurrent neural network was trained to abstract the particular bias of each individual multiobjective algorithm, generating a heuristic that could be useful to predict some of the relevant aspects of the three-dimensional conformation process formation, known as protein folding.

Keywords: Ab initio heuristic modeling, multiobjective optimization, protein structure prediction, recurrent neural network

Procedia PDF Downloads 177

Authors: Ksenia Volkova, Artur Petrosyan, Ignatii Dubyshkin, Alexei Ossadtchi

Abstract:

Brain-computer interfaces are a growing research field producing many implementations that find use in different fields and are used for research and practical purposes. Despite the popularity of the implementations using non-invasive neuroimaging methods, radical improvement of the state channel bandwidth and, thus, decoding accuracy is only possible by using invasive techniques. Electrocorticography (ECoG) is a minimally invasive neuroimaging method that provides highly informative brain activity signals, effective analysis of which requires the use of machine learning methods that are able to learn representations of complex patterns. Deep learning is a family of machine learning algorithms that allow learning representations of data with multiple levels of abstraction. This study explores the potential of deep learning approaches for ECoG processing, decoding movement intentions and the perception of proprioceptive information. To obtain synchronous recording of kinematic movement characteristics and corresponding electrical brain activity, a series of experiments were carried out, during which subjects performed finger movements at their own pace. Finger movements were recorded with a three-axis accelerometer, while ECoG was synchronously registered from the electrode strips that were implanted over the contralateral sensorimotor cortex. Then, multichannel ECoG signals were used to track finger movement trajectory characterized by accelerometer signal. This process was carried out both causally and non-causally, using different position of the ECoG data segment with respect to the accelerometer data stream. The recorded data was split into training and testing sets, containing continuous non-overlapping fragments of the multichannel ECoG. A deep convolutional neural network was implemented and trained, using 1-second segments of ECoG data from the training dataset as input. To assess the decoding accuracy, correlation coefficient r between the output of the model and the accelerometer readings was computed. After optimization of hyperparameters and training, the deep learning model allowed reasonably accurate causal decoding of finger movement with correlation coefficient r = 0.8. In contrast, the classical Wiener-filter like approach was able to achieve only 0.56 in the causal decoding mode. In the noncausal case, the traditional approach reached the accuracy of r = 0.69, which may be due to the presence of additional proprioceptive information. This result demonstrates that the deep neural network was able to effectively find a representation of the complex top-down information related to the actual movement rather than proprioception. The sensitivity analysis shows physiologically plausible pictures of the extent to which individual features (channel, wavelet subband) are utilized during the decoding procedure. In conclusion, the results of this study have demonstrated that a combination of a minimally invasive neuroimaging technique such as ECoG and advanced machine learning approaches allows decoding motion with high accuracy. Such setup provides means for control of devices with a large number of degrees of freedom as well as exploratory studies of the complex neural processes underlying movement execution.

Keywords: brain-computer interface, deep learning, ECoG, movement decoding, sensorimotor cortex

Procedia PDF Downloads 132