Afef SAIDI

Followers

Following

Co-author

Public Views

Has a Ph.D. in Electrical Engineering at the University of National Engineering School of Carthage. She is a research member at Laboratory of Advanced Systems of Polytechnic School of Tunisia. She has a master degree in Biophysics and Medical Imaging in Higher Institute of Medical Technologies of Tunis. Since 2018, she started as teaching assistant in the Higher Institute of Information and Communication Technologies of Tunisia at the Department of Industrial Information Technology. Her research centers on the solutions for the implementation of classification techniques based on parallel architectures and reconfigurable platforms. She can be contacted at email: [email protected]

less

Uploads

articles by Afef SAIDI

Accelerating Deep Neural Networks implementation: A survey

by Mariem Dhouibi and Afef SAIDI

IET Computers & Digital Techniques, 2021

Abstract Recently, Deep Learning (DL) applications are getting more and more involved in differen... more Abstract Recently, Deep Learning (DL) applications are getting more and more involved in different fields. Deploying such Deep Neural Networks (DNN) on embedded devices is still a challenging task considering the massive requirement of computation and storage. Given that the number of operations and parameters increases with the complexity of the model architecture, the performance will strongly depend on the hardware target resources and basically the memory footprint of the accelerator. Recent research studies have discussed the benefit of implementing some complex DL applications based on different models and platforms. However, it is necessary to guarantee the best performance when designing hardware accelerators for DL applications to run at full speed, despite the constraints of low power, high accuracy and throughput. Field Programmable Gate Arrays (FPGAs) are promising platforms for the deployment of large-scale DNN which seek to reach a balance between the above objectives. Besides, the growing complexity of DL models has made researches think about applying optimization techniques to make them more hardware-friendly. Herein, DL concept is presented. Then, a detailed description of different optimization techniques used in recent research works is explored. Finally, a survey of research works aiming to accelerate the implementation of DNN models on FPGAs is provided.

Papers by Afef SAIDI

Acceleration of convolutional neural network based diabetic retinopathy diagnosis system on field programmable gate array

International Journal of Informatics and Communication Technology (IJ-ICT)

Diabetic retinopathy (DR) is one of the most common causes of blindness. The necessity for a robu... more Diabetic retinopathy (DR) is one of the most common causes of blindness. The necessity for a robust and automated DR screening system for regular examination has long been recognized in order to identify DR at an early stage. In this paper, an embedded DR diagnosis system based on convolutional neural networks (CNNs) has been proposed to assess the proper stage of DR. We coupled the power of CNN with transfer learning to design our model based on state-of-the-art architecture. We preprocessed the input data, which is color fundus photography, to reduce undesirable noise in the image. After training many models on the dataset, we chose the adopted ResNet50 because it produced the best results, with a 92.90% accuracy. Extensive experiments and comparisons with other research work show that the proposed method is effective. Furthermore, the CNN model has been implemented on an embedded target to be a part of a medical instrument diagnostic system. We have accelerated our model inferenc...

Download

CNN inference acceleration on limited resources FPGA platforms_epilepsy detection case study

International Journal of Informatics and Communication Technology (IJ-ICT)

The use of a convolutional neural network (CNN) to analyze and classify electroencephalogram (EEG... more The use of a convolutional neural network (CNN) to analyze and classify electroencephalogram (EEG) signals has recently attracted the interest of researchers to identify epileptic seizures. This success has come with an enormous increase in the computational complexity and memory requirements of CNNs. For the sake of boosting the performance of CNN inference, several hardware accelerators have been proposed. The high performance and flexibility of the field programmable gate array (FPGA) make it an efficient accelerator for CNNs. Nevertheless, for resource-limited platforms, the deployment of CNN models poses significant challenges. For an ease of CNN implementation on such platforms, several tools and frameworks have been made available by the research community along with different optimization techniques. In this paper, we proposed an FPGA implementation for an automatic seizure detection approach using two CNN models, namely VGG-16 and ResNet-50. To reduce the model size and com...

Download

Hybrid CNN-SVM classifier for efficient depression detection system

2020 4th International Conference on Advanced Systems and Emergent Technologies (IC_ASET), 2020

Depression is a serious debilitating mental disorder affecting people from all ages all over the ... more Depression is a serious debilitating mental disorder affecting people from all ages all over the world. The number of depression cases increases annually in a continuous way. Due to the complexity of traditional techniques based on clinical diagnosis, there is a need for an automatic detection system of the depression. In this paper we present a novel audio-based approach to automatically detect depression using hybrid model. This model combines convolutional neural networks (CNN) and support vector machines (SVM), where SVM takes the place of the fully connected layers in CNN. In this proposed model, the features are automatically extracted using CNN and the classification is done using the SVM classifier. This approach was evaluated using DAIC-WOZ dataset provided by AVEC 2016 depression analysis sub-challenge. Experimental results showed that our hybrid model achieved an accuracy of 68% which outperform the CNN model (58.57%). Compared to the previous audio-based works using the same DAIC-WOZ dataset, our work showed a significant improvement in terms of accuracy, precision and recall.

FPGA-based implementation of classification techniques: A survey

Integration, 2021

Recently, a number of classification techniques have been introduced. However, processing large d... more Recently, a number of classification techniques have been introduced. However, processing large dataset in a reasonable time has become a major challenge. This made classification task more complex and expensive in calculation. Thus, the need for solutions to overcome these constraints such as field programmable gate arrays (FPGAs). In this paper, we give an overview of the various classification techniques. Then, we present the existing FPGA based implementation of these classification methods. After that, we investigate the confronted challenges and the optimizations strategies. Finally, we highlight the hardware accelerator architectures and tools for hardware design suggested to improve the FPGA implementation of classification methods.

A novel epileptic seizure detection system using scalp EEG signals based on hybrid CNN-SVM classifier

2021 IEEE Symposium on Industrial Electronics & Applications (ISIEA), 2021

Epilepsy is a neurological disorder that affects more than 2% of the world’s population. Encephal... more Epilepsy is a neurological disorder that affects more than 2% of the world’s population. Encephalography (EEG) is a commonly clinical tool used for the diagnosis of epilepsy. However, traditional approaches based on visual inspection of EEG signals are tedious and complex. Thus, several automatic seizure detection approaches based on machine learning techniques have been proposed. In this study, a hybrid model for the detection of epileptic seizure is proposed, where convolutional neural network (CNN) is used for automatic feature extraction of EEG signals and support vector machines (SVM) is used for epileptic seizure classification. The proposed approach was evaluated using the Children’s Hospital Boston-Massachusetts Institute of Technology (CHB-MIT) dataset. Experimental results showed that the accuracy of the combined CNN-SVM model outperforms the CNN baseline model. The proposed approach provides a substantial increase in seizure prediction performance in terms of sensitivity compared to both classical machine learning approaches and CNN model that have been presented in the previous studies.

FPGA implementation of EEG signal analysis system for the detection of epileptic seizure

2018 International Conference on Advanced Systems and Electric Technologies (IC_ASET)

Epilepsy is a neurological disorder identified by successive sudden seizures resulting from trans... more Epilepsy is a neurological disorder identified by successive sudden seizures resulting from transient and recurrent electrical discharges in the brain. Their detection is possible by analyzing electroencephalogram (EEG) signals. EEG is the commonly used clinical tool for neurological disorders due to the valuable information that contains about brain activities. The work presented in this paper is oriented towards two parts. A first part is dedicated to the description and the validation of our detection system using the classification algorithm, AntMiner+. The second part focuses on the design and the implementation of our application on the Xilinx ZC702 embedded board.

Afef SAIDI

Uploads

articles by Afef SAIDI

Papers by Afef SAIDI

Log In