An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet
Convolutional Neural Networks (CNN) continue to dominate research in the area of hardware acceleration using Field Programmable Gate Arrays (FPGA), proving its effectiveness in a variety of computer vision applications such as object segmentation, image classification, face detection, and traffic si...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-09-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/10/18/2272 |
| _version_ | 1797519553987608576 |
|---|---|
| author | Safa Bouguezzi Hana Ben Fredj Tarek Belabed Carlos Valderrama Hassene Faiedh Chokri Souani |
| author_facet | Safa Bouguezzi Hana Ben Fredj Tarek Belabed Carlos Valderrama Hassene Faiedh Chokri Souani |
| author_sort | Safa Bouguezzi |
| collection | DOAJ |
| description | Convolutional Neural Networks (CNN) continue to dominate research in the area of hardware acceleration using Field Programmable Gate Arrays (FPGA), proving its effectiveness in a variety of computer vision applications such as object segmentation, image classification, face detection, and traffic signs recognition, among others. However, there are numerous constraints for deploying CNNs on FPGA, including limited on-chip memory, CNN size, and configuration parameters. This paper introduces Ad-MobileNet, an advanced CNN model inspired by the baseline MobileNet model. The proposed model uses an Ad-depth engine, which is an improved version of the depth-wise separable convolution unit. Moreover, we propose an FPGA-based implementation model that supports the Mish, TanhExp, and ReLU activation functions. The experimental results using the CIFAR-10 dataset show that our Ad-MobileNet has a classification accuracy of 88.76% while requiring little computational hardware resources. Compared to state-of-the-art methods, our proposed method has a fairly high recognition rate while using fewer computational hardware resources. Indeed, the proposed model helps to reduce hardware resources by more than 41% compared to that of the baseline model. |
| first_indexed | 2024-03-10T07:44:24Z |
| format | Article |
| id | doaj.art-5ec2800df62e40d885a430dc19359d47 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T07:44:24Z |
| publishDate | 2021-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-5ec2800df62e40d885a430dc19359d472023-11-22T12:48:30ZengMDPI AGElectronics2079-92922021-09-011018227210.3390/electronics10182272An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNetSafa Bouguezzi0Hana Ben Fredj1Tarek Belabed2Carlos Valderrama3Hassene Faiedh4Chokri Souani5Laboratoire de Microélectronique et Instrumentation, Faculté des Sciences de Monastir, Université de Monastir, Monastir 5019, TunisiaLaboratoire de Microélectronique et Instrumentation, Faculté des Sciences de Monastir, Université de Monastir, Monastir 5019, TunisiaLaboratoire de Microélectronique et Instrumentation, Faculté des Sciences de Monastir, Université de Monastir, Monastir 5019, TunisiaElectronics and Microelectronics Unit (SEMi), University of Mons, 7000 Mons, BelgiumInstitut Supérieur des Sciences Appliquées et de Technologie de Sousse, Université de Sousse, Sousse 4003, TunisiaInstitut Supérieur des Sciences Appliquées et de Technologie de Sousse, Université de Sousse, Sousse 4003, TunisiaConvolutional Neural Networks (CNN) continue to dominate research in the area of hardware acceleration using Field Programmable Gate Arrays (FPGA), proving its effectiveness in a variety of computer vision applications such as object segmentation, image classification, face detection, and traffic signs recognition, among others. However, there are numerous constraints for deploying CNNs on FPGA, including limited on-chip memory, CNN size, and configuration parameters. This paper introduces Ad-MobileNet, an advanced CNN model inspired by the baseline MobileNet model. The proposed model uses an Ad-depth engine, which is an improved version of the depth-wise separable convolution unit. Moreover, we propose an FPGA-based implementation model that supports the Mish, TanhExp, and ReLU activation functions. The experimental results using the CIFAR-10 dataset show that our Ad-MobileNet has a classification accuracy of 88.76% while requiring little computational hardware resources. Compared to state-of-the-art methods, our proposed method has a fairly high recognition rate while using fewer computational hardware resources. Indeed, the proposed model helps to reduce hardware resources by more than 41% compared to that of the baseline model.https://www.mdpi.com/2079-9292/10/18/2272FPGAMobileNetdepthwise separable convolutionCNNdeep learning |
| spellingShingle | Safa Bouguezzi Hana Ben Fredj Tarek Belabed Carlos Valderrama Hassene Faiedh Chokri Souani An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet Electronics FPGA MobileNet depthwise separable convolution CNN deep learning |
| title | An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet |
| title_full | An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet |
| title_fullStr | An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet |
| title_full_unstemmed | An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet |
| title_short | An Efficient FPGA-Based Convolutional Neural Network for Classification: Ad-MobileNet |
| title_sort | efficient fpga based convolutional neural network for classification ad mobilenet |
| topic | FPGA MobileNet depthwise separable convolution CNN deep learning |
| url | https://www.mdpi.com/2079-9292/10/18/2272 |
| work_keys_str_mv | AT safabouguezzi anefficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT hanabenfredj anefficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT tarekbelabed anefficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT carlosvalderrama anefficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT hassenefaiedh anefficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT chokrisouani anefficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT safabouguezzi efficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT hanabenfredj efficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT tarekbelabed efficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT carlosvalderrama efficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT hassenefaiedh efficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet AT chokrisouani efficientfpgabasedconvolutionalneuralnetworkforclassificationadmobilenet |