FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system
In this presentation, we report the results of applying a binarised Convolutional Neural Network (CNN) and a Field Programmable Gate Array (FPGA) for image-based object recognition. While the demand rises for robots with robust object recognition implemented with Neural Networks, a trade-off between...
| Main Authors: | , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2017
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/64363/1/E%26NT%20poster%20121117%201.pdf |
| _version_ | 1796977883105722368 |
|---|---|
| author | Shuto, Daisuke Abbas, Z. Sulaiman, N. Tamukoh, H. |
| author_facet | Shuto, Daisuke Abbas, Z. Sulaiman, N. Tamukoh, H. |
| author_sort | Shuto, Daisuke |
| collection | UPM |
| description | In this presentation, we report the results of applying a binarised Convolutional Neural Network (CNN) and a Field Programmable Gate Array (FPGA) for image-based object recognition. While the demand rises for robots with robust object recognition implemented with Neural Networks, a trade-off between data processing rate and power consumption persists. Some applications utilise GPGPU (General Purpose computing on Graphics Processing Units), which results in high power consumption thus undesirable for embedded systems, while the others communicate with cloud computers to minimise computational resources at the clients’ side, i.e. robots, raising another concern that the robots are unable to perform object recognition without the servers and network connections. To overcome these difficulties, we propose an embedded object recognition system implemented with a binarised CNN and an FPGA. FPGAs consist of a matrix of reconfigurable logic gates allowing parallel computing which befits most image processing algorithms such as the CNN. We train the binarised CNN on one of our datasets that contain images of several kinds of food and beverages. The results of the experiments show that the binarised CNN with an FPGA maintains high accuracy as well as real-time computation, suggesting that the proposed system is suitable for robots to perform their tasks in a real-world environment without needing to communicate with a server. |
| first_indexed | 2024-03-06T09:46:34Z |
| format | Conference or Workshop Item |
| id | upm.eprints-64363 |
| institution | Universiti Putra Malaysia |
| language | English |
| last_indexed | 2024-03-06T09:46:34Z |
| publishDate | 2017 |
| record_format | dspace |
| spelling | upm.eprints-643632018-07-04T02:38:11Z http://psasir.upm.edu.my/id/eprint/64363/ FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system Shuto, Daisuke Abbas, Z. Sulaiman, N. Tamukoh, H. In this presentation, we report the results of applying a binarised Convolutional Neural Network (CNN) and a Field Programmable Gate Array (FPGA) for image-based object recognition. While the demand rises for robots with robust object recognition implemented with Neural Networks, a trade-off between data processing rate and power consumption persists. Some applications utilise GPGPU (General Purpose computing on Graphics Processing Units), which results in high power consumption thus undesirable for embedded systems, while the others communicate with cloud computers to minimise computational resources at the clients’ side, i.e. robots, raising another concern that the robots are unable to perform object recognition without the servers and network connections. To overcome these difficulties, we propose an embedded object recognition system implemented with a binarised CNN and an FPGA. FPGAs consist of a matrix of reconfigurable logic gates allowing parallel computing which befits most image processing algorithms such as the CNN. We train the binarised CNN on one of our datasets that contain images of several kinds of food and beverages. The results of the experiments show that the binarised CNN with an FPGA maintains high accuracy as well as real-time computation, suggesting that the proposed system is suitable for robots to perform their tasks in a real-world environment without needing to communicate with a server. 2017 Conference or Workshop Item PeerReviewed text en http://psasir.upm.edu.my/id/eprint/64363/1/E%26NT%20poster%20121117%201.pdf Shuto, Daisuke and Abbas, Z. and Sulaiman, N. and Tamukoh, H. (2017) FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system. In: 5th International Symposium on Applied Engineering and Sciences (SAES2017), 14-15 Nov. 2017, Universiti Putra Malaysia. (p. 7). |
| spellingShingle | Shuto, Daisuke Abbas, Z. Sulaiman, N. Tamukoh, H. FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system |
| title | FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system |
| title_full | FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system |
| title_fullStr | FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system |
| title_full_unstemmed | FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system |
| title_short | FPGA-enabled binarised convolutional neural networks toward real-time embedded object recognition system |
| title_sort | fpga enabled binarised convolutional neural networks toward real time embedded object recognition system |
| url | http://psasir.upm.edu.my/id/eprint/64363/1/E%26NT%20poster%20121117%201.pdf |
| work_keys_str_mv | AT shutodaisuke fpgaenabledbinarisedconvolutionalneuralnetworkstowardrealtimeembeddedobjectrecognitionsystem AT abbasz fpgaenabledbinarisedconvolutionalneuralnetworkstowardrealtimeembeddedobjectrecognitionsystem AT sulaimann fpgaenabledbinarisedconvolutionalneuralnetworkstowardrealtimeembeddedobjectrecognitionsystem AT tamukohh fpgaenabledbinarisedconvolutionalneuralnetworkstowardrealtimeembeddedobjectrecognitionsystem |