Computer
Permanent URI for this collectionhttps://dspace.univ-boumerdes.dz/handle/123456789/3082
Browse
7 results
Search Results
Item Action detection using deep learning shoplifting detection framework(Université M’hamed Bougara de Boumerdes : Institut de Genie Electrique et Electronique, 2024) Bettahar, Mohammed Nadir; Touzout, Walid ( Supervisor)This project delves into the application of deep learning for action detection, with a specifi cfocu so nidentifyin gshopliftin gbehavior si nretai lenvironments. Th egrowing need for automated surveillance systems that can efficient lya ndaccurate lydetect suspicious activities has motivated this work. Shoplifting action detection is the process of identifying and localizing shoplifting activities in a video by findin gbot hwhere and when an action occurs within a video clip and determining what action is being performed. A key challenge lies in preparing a dataset that reflect sth ecomplexity of real-world scenarios, which was addressed by employing semi-supervised learning techniques. The use of You Only Watch Once version 9 (YOLOv9) object detection model,its tracking function, was instrumental in the automation of labeling and tracking objects within the shoplifting video dataset, ensuring a reliable foundation for action detection. To evaluate the effectivenes so fth esystem ,th eYo uOnl yWatc hOnc eversio n2 (YOWOv2) model was used, conducting comprehensive training and testing across a variety of shoplifting situations. This allowed for a detailed assessment of the model’s ability to recognize and generalize diverse shoplifting actions, even in challenging environments. The results show that the models can detect suspicious behavior, offerin ga promising tool for improving retail security. This work contributes to the broader field of shoplifting detection by providing insights into how deep learning techniques can enhance real-time surveillance and reduce theft in retail settings, with potential applications in other domains of anomaly detection. The YOWOv2-Medium-16-frames model gave the best performance with 54.74% frame mean average precision and 42.67% in video mean average precision.Item Atrial fibrillation analysis by deep learning.(Université M’hamed Bougara de Boumerdes : Institut de Genie Electrique et Electronique, 2024) Agli, Wafa; Daamouche, Abdelhamid (Supervisor)Atrial fibrillation (AF), an increasingly prevalent cardia carrhythmia, is a major contributor to stroke, heart failure, and premature mortality. Traditional manual screening for AF using electrocardiography (ECG) is not only time-consuming but also susceptible to human error, underscoring the urgent need for automated diagnostic tools. This study addresses this challenge by developing advanced computer-aided diagnostic methods leveraging deep learning for the automatic detection of AF. We introduce innovative one-dimensional (1D) and two-dimensional (2D) convolutional neural network (CNN) models specifically designed for the precise classification of ECG signals into normal or atrial fibrillation categories. Our methodology includes a meticulous preprocessing phase where each ECG record is filtere dan dpeak sare accurately detected using the XQRS algorithm. The signals are then segmented into beats with an 80-sample window, which serve as critical features for subsequent classification. The extracted features are fed into our CNN architectures for classification. The models are trained and evaluated using the MIT-BIH Atrial Fibrillation Database, and their generalization capability is further validated with unseen data from the PhysioNet/Computing in Cardiology Challenge 2017 database, following an inter-subject approach. To enhance the robustness of our models, we employ data augmentation techniques. Our comprehensive evaluation demonstrates that the 1D-CNN model achieves a remarkable total accuracy of 95% and an F1 score of 96.81%, while the 2D-CNN model attains an exceptional accuracy and F1 score of 99.84%. These results underscore the efficacy of our approach in accurately classifying ECG signals and highlight the potential of our models for real-world clinical applications, offering a substantial improvement in AF screening processes.Item Deep learning methods for speech synthesis.(Université M’hamed Bougara de Boumerdes : Institut de Genie Electrique et Electronique, 2023) Bechetella, Abderraouf; Tabet, Youcef (Supervisor)In the name of Allah, the Most Merciful and the Most Gracious, We would like to begin by expressing our gratitude to Allah for His blessings, guidance, and unwavering support throughout this research journey. His mercy and wisdom have been a source of strength and inspiration. We would also like to extend our heartfelt appreciation to our supervisor, Dr. Tabet, for his invaluable guidance, mentor-ship, and continuous encouragement. His expertise and wisdom have been instrumental in shaping the direction of this project. We would like to acknowledge and appreciate the teachers and staff of The Institute of Electrical and Electronics Engineering for their knowledge, guidance, and support throughout our academic journey. Their dedication and expertise have greatly contributed to our growth and learning. Lastly, We would like to express our sincere gratitude to our family, friends, and loved ones for their unwavering belief in us, their love, and their constant support. Their encouragement and presence have been a source of motivation and strength.Item Diabetic retinopathy grading using deep neural networks(Université M’hamed Bougara de Boumerdes : Institut de Genie Electrique et Electronique, 2023) Temmam, Amira; Ahmed Gaid, Chaima; Daamouche, Abdelhamid (Supervisor)Diabetic Retinopathy (DR) is a chronic disease and the leading cause of blindness and visual impairment among diabetic patients making early detection and classificatio nof DR crucial for effectiv etreatment .Thi sprojec tutilize sstate-of-the-ar tconvolutiona lneu- ral networks and transfer learning techniques to analyze retinal images and classify the severity of the disease on a scale of 0 to 4 (ranging from healthy to proliferative). We employ architectures such as EfficientNetB 1,InceptionV 3,Xceptio n,a ndMobileNet V2to achieve accurate classification .Throug h aserie so fexperiment san devaluations ,ou rsys- tem achieves an overall accuracy of 80.0% in classifying DR images. The results showcase the significan tpotentia lo fdee plearnin gi nassistin ghealthcar eprofessional swit hearly diagnosis and treatment planning, thereby improving patient well-being.Item EMG signals classification for neuromuscular diseases detectionusing deep learning(Université M’hamed Bougara de Boumerdes : Institut de Genie Electrique et Electronique, 2023) Loubar, Lidia; Toubal, Maria; Boutellaa, Elhocine (Supervisor)Neuromuscular diseases are particular impairments that affect the muscle tissue or nervous system part connected to muscles. Electromyography (EMG) signals are valuable biosignals for the diagnosis of neuromuscular diseases. However, the classification of EMG signals is a challenging task due to the complexity of the signals and the variability of the diseases. In this project, we address the problem of EMG signals classification for the detection of neuromuscular diseases using deep learning techniques. The main goal of our project is to develop a robust deep-learning model that performs well on unseen data, thereby improving the reliability of diagnosis in real-life scenarios. To achieve this, we design a model which we train and evaluate on a dataset of EMG signals from patients with different neuromuscular diseases. We assess the performance of our designed model using two different methods : the train-test split approach, commonly employed in the existing literature, and the subject-independent evaluation method, which ensures that the model is tested on completely unseen data. The results show that the model achieves excellent performance on the train-test split approach. However, the second method produces varied and uneven scores for different patients, suggesting that EMG data of certain individuals may be more challenging to classify accurately. Nonetheless, some patients exhibit highly accurate classifications, demonstrating the potential performance of our designed model. The obtained results indicate the potential of the developed tool for the diagnosis of neuromuscular diseases.Item Brain tumor classificaion using deep learning.(2022) Berrichi, Ryad; Namane, Rachid (Supervisor)Brain tumors are a common type of cancer that affects brain tissue. They often cause symp- toms such as headaches or seizures. They are usually diagnosed through brain scans such as magnetic resonance imaging (MRI). In recent years, computer scientists have developed algorithms that have shown promising results in automatically classifying these images into various types using deep learning models, which is a type of machine learning that uses artificial neural networks to recognize patterns in data. Publicly available MRI scans (1500 cancerous and 1500 non-cancerous) are used to train deep learning models: VGG16, VGG19, ResNet50, and Xception. Each model is implemented using three approaches, namely: implementation from scratch, transfer learning, and fine- tuning. This comparative study aims to find the best approach for training models on small datasets. The obtained overall accuracies ranged from 88% to 99%.Item Deep learning-based mobile application for plant disease diagnosis.(2022) Beggar, Ikram; Daamouche, Abdelhamid (Supervisor)Plants are susceptible to a variety of diseases in farming. The impact of sudden climatic change harms their growth causing dangerous viruses and pests. Plant diseases are one of the most serious problems confronting agriculture across the world and harming the health, economy, and livelihood of the human population. The majority of traditional plant pests' diagnosis methods rely on human visual observation and inspection. However, this approach is time-consuming and requires strong agricultural skills and significant human efforts. Recent breakthroughs in computer vision and Deep Learning provide a potential pathway for developing a plant disease diagnosis system that will be able to detect plant diseases in different geographical regions with fewer human interventions. This dissertation presents a Deep Learning powered mobile-based system to automate the early identification of plant diseases. The developed system uses Convolutional Neural Networks to classify plant leaves into 38 classes with 14 types of plants. To train and test our models, we used two datasets from Kaggle containing about 87000 and 20803 images of healthy and diseased plant leaves respectively. After the training, we assessed our models using some classification evaluation metrics such as the accuracy and f1-score and found that they were able to correctly classify most of the images in the test set. Finally, to increase the usability of our algorithm, we developed a smartphone application that runs on both Android and iOS operating systems with a simple user interface using the Flutter framework. This application allows farmers to capture pictures of their infected plant leaves or import them from their phone library and then it displays the disease category with the plant name and the accuracy of the prediction. This approach is supposed to provide farmers with a better opportunity to maintain their crops' health and minimize the use the incorrect fertilizers that might damage the plants and the environment.