Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission
Deep learning-based image recognition systems have rapidly evolved. Due to the extensive processing load of the deep neural network (DNN) on graphic processing units (GPUs), the DNN model is deployed on the cloud server. Images or videos are forwarded from user terminals through the network to the s...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9869327/ |
| _version_ | 1798032030550720512 |
|---|---|
| author | Mutsuki Nakahara Mai Nishimura Yoshitaka Ushiku Takayuki Nishio Kazuki Maruta Yu Nakayama Daisuke Hisano |
| author_facet | Mutsuki Nakahara Mai Nishimura Yoshitaka Ushiku Takayuki Nishio Kazuki Maruta Yu Nakayama Daisuke Hisano |
| author_sort | Mutsuki Nakahara |
| collection | DOAJ |
| description | Deep learning-based image recognition systems have rapidly evolved. Due to the extensive processing load of the deep neural network (DNN) on graphic processing units (GPUs), the DNN model is deployed on the cloud server. Images or videos are forwarded from user terminals through the network to the server. In recent years, edge computing has gained popularity as a means of reducing the data traffic in the backbone network. However, the last one-mile access network between an edge server and user terminals will still be congested because a large amount of data such as video/image files must be forwarded. In particular, when computer vision applications such as image recognition are loaded in the edge network, a large amount of data is forwarded although the edge server always may not need the high-definition image. This paper proposes an image compression and progressive retransmission scheme for deep learning-based image recognition systems to reduce image data traffic and alleviate network congestion. The proposed method introduces an entropy-based threshold calculated from posterior probabilities from a deep learning model’s output layer. Entropy is an extremely effective metric because it can be used as an indicator independent of the number of classification labels in the DNN model. The thresholding can control the image retransmission and reduce traffic while maintaining image recognition accuracy. We implement the proposed scheme on the edge server and reveal the relationship between the data compression and the recognition accuracy through simulation evaluation. As a result, we indicate that an entropy-based threshold reduces the overall ambiguity of the accuracy of image recognition. Moreover, when a higher accuracy recognition model with more accuracy is combined with a retransmission scheme, it becomes the more effective. |
| first_indexed | 2024-04-11T20:06:10Z |
| format | Article |
| id | doaj.art-bb0680dbc3d642718d71f1bd40105df3 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-11T20:06:10Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-bb0680dbc3d642718d71f1bd40105df32022-12-22T04:05:18ZengIEEEIEEE Access2169-35362022-01-0110912539126210.1109/ACCESS.2022.32021729869327Edge Computing-Assisted DNN Image Recognition System With Progressive Image RetransmissionMutsuki Nakahara0https://orcid.org/0000-0002-7186-4645Mai Nishimura1Yoshitaka Ushiku2https://orcid.org/0000-0002-9014-1389Takayuki Nishio3https://orcid.org/0000-0003-1026-319XKazuki Maruta4https://orcid.org/0000-0001-9635-3182Yu Nakayama5https://orcid.org/0000-0002-6945-7055Daisuke Hisano6https://orcid.org/0000-0002-5766-3812Graduate School of Engineering, Osaka University, Suita, Osaka, JapanOMRON SINIC X Corporation, Bunkyo, Tokyo, JapanOMRON SINIC X Corporation, Bunkyo, Tokyo, JapanSchool of Engineering, Tokyo Institute of Technology, Ookayama, Meguro, Tokyo, JapanDepartment of Electrical Engineering, Tokyo University of Science, Katsushika, Tokyo, JapanInstitute of Engineering, Tokyo University of Agriculture and Technology, Koganei, Tokyo, JapanGraduate School of Engineering, Osaka University, Suita, Osaka, JapanDeep learning-based image recognition systems have rapidly evolved. Due to the extensive processing load of the deep neural network (DNN) on graphic processing units (GPUs), the DNN model is deployed on the cloud server. Images or videos are forwarded from user terminals through the network to the server. In recent years, edge computing has gained popularity as a means of reducing the data traffic in the backbone network. However, the last one-mile access network between an edge server and user terminals will still be congested because a large amount of data such as video/image files must be forwarded. In particular, when computer vision applications such as image recognition are loaded in the edge network, a large amount of data is forwarded although the edge server always may not need the high-definition image. This paper proposes an image compression and progressive retransmission scheme for deep learning-based image recognition systems to reduce image data traffic and alleviate network congestion. The proposed method introduces an entropy-based threshold calculated from posterior probabilities from a deep learning model’s output layer. Entropy is an extremely effective metric because it can be used as an indicator independent of the number of classification labels in the DNN model. The thresholding can control the image retransmission and reduce traffic while maintaining image recognition accuracy. We implement the proposed scheme on the edge server and reveal the relationship between the data compression and the recognition accuracy through simulation evaluation. As a result, we indicate that an entropy-based threshold reduces the overall ambiguity of the accuracy of image recognition. Moreover, when a higher accuracy recognition model with more accuracy is combined with a retransmission scheme, it becomes the more effective.https://ieeexplore.ieee.org/document/9869327/Edge computingcomputer visionimage recognitionretransmission system |
| spellingShingle | Mutsuki Nakahara Mai Nishimura Yoshitaka Ushiku Takayuki Nishio Kazuki Maruta Yu Nakayama Daisuke Hisano Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission IEEE Access Edge computing computer vision image recognition retransmission system |
| title | Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission |
| title_full | Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission |
| title_fullStr | Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission |
| title_full_unstemmed | Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission |
| title_short | Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission |
| title_sort | edge computing assisted dnn image recognition system with progressive image retransmission |
| topic | Edge computing computer vision image recognition retransmission system |
| url | https://ieeexplore.ieee.org/document/9869327/ |
| work_keys_str_mv | AT mutsukinakahara edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission AT mainishimura edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission AT yoshitakaushiku edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission AT takayukinishio edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission AT kazukimaruta edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission AT yunakayama edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission AT daisukehisano edgecomputingassisteddnnimagerecognitionsystemwithprogressiveimageretransmission |