The 2020 Low-Power Computer Vision Challenge

Xiao Hu; Ming Ching Chang; Yuwei Chen; Rahul Sridhar; Zhenyu Hu; Yunhe Xue; Zhenyu Wu; Pengcheng Pi; Jiayi Shen; Jianchao Tan; Xiangru Lian; Ji Liu; Zhangyang Wang; Chia Hsiang Liu; Yu Shin Han; Yuan Yao Sung; Yi Lee; Kai Chiang Wu; Wei Xiang Guo; Rick Lee; Shengwen Liang; Zerun Wang; Guiguang Ding; Gang Zhang; Teng Xi; Yubei Chen; Han Cai; Ligeng Zhu; Zhekai Zhang; Song Han; Seonghwan Jeong; Youngmin Kwon; Tianzhe Wang; Jeffery Pan

doi:10.1109/AICAS51828.2021.9458522

The 2020 Low-Power Computer Vision Challenge

Xiao Hu, Ming Ching Chang, Yuwei Chen, Rahul Sridhar, Zhenyu Hu, Yunhe Xue, Zhenyu Wu, Pengcheng Pi, Jiayi Shen, Jianchao Tan, Xiangru Lian, Ji Liu, Zhangyang Wang, Chia Hsiang Liu, Yu Shin Han, Yuan Yao Sung, Yi Lee, Kai Chiang Wu, Wei Xiang Guo, Rick LeeShengwen Liang, Zerun Wang, Guiguang Ding, Gang Zhang, Teng Xi, Yubei Chen, Han Cai, Ligeng Zhu, Zhekai Zhang, Song Han, Seonghwan Jeong, Youngmin Kwon, Tianzhe Wang, Jeffery Pan

研究成果: Conference contribution › 同行評審

3 引文斯高帕斯（Scopus）

摘要

AI computer vision has advanced significantly in recent years. IoT and edge computing devices such as mobile phones have become the primary computing platform for many end users. Mobile devices such as robots and drones that rely on batteries demand for energy efficient computation. Since 2015, the IEEE Annual International Low-Power Computer Vision Challenge (LPCVC) was held to identify energy-efficient AI and computer vision solutions. The 2020 LPCVC includes three challenge tracks: (1) PyTorch UAV Video Track, (2) FPGA Image Track, and (3) On-device Visual Intelligence Competition (OVIC) Tenforflow Track. This paper summarizes the 2020 winning solutions from the three tracks of LPCVC competitions. Methods and future directions for energy-efficient AI and computer vision research are discussed.

原文	English
主出版物標題	2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021
發行者	Institute of Electrical and Electronics Engineers Inc.
ISBN（電子）	9781665419130
DOIs	https://doi.org/10.1109/AICAS51828.2021.9458522
出版狀態	Published - 6 6月 2021
事件	3rd IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021 - Washington, 美國持續時間: 6 6月 2021 → 9 6月 2021

出版系列

名字	2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021

Conference

Conference	3rd IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021
國家/地區	美國
城市	Washington
期間	6/06/21 → 9/06/21

存取文件

10.1109/AICAS51828.2021.9458522

其它文件與鏈接

Link to publication in Scopus

引用此

Hu, X., Chang, M. C., Chen, Y., Sridhar, R., Hu, Z., Xue, Y., Wu, Z., Pi, P., Shen, J., Tan, J., Lian, X., Liu, J., Wang, Z., Liu, C. H., Han, Y. S., Sung, Y. Y., Lee, Y., Wu, K. C., Guo, W. X., ... Pan, J. (2021). The 2020 Low-Power Computer Vision Challenge. 於 2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021 文章 9458522 (2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/AICAS51828.2021.9458522

@inproceedings{2ecb387c73c74f239bd2537d86fdfb61,

title = "The 2020 Low-Power Computer Vision Challenge",

abstract = "AI computer vision has advanced significantly in recent years. IoT and edge computing devices such as mobile phones have become the primary computing platform for many end users. Mobile devices such as robots and drones that rely on batteries demand for energy efficient computation. Since 2015, the IEEE Annual International Low-Power Computer Vision Challenge (LPCVC) was held to identify energy-efficient AI and computer vision solutions. The 2020 LPCVC includes three challenge tracks: (1) PyTorch UAV Video Track, (2) FPGA Image Track, and (3) On-device Visual Intelligence Competition (OVIC) Tenforflow Track. This paper summarizes the 2020 winning solutions from the three tracks of LPCVC competitions. Methods and future directions for energy-efficient AI and computer vision research are discussed.",

keywords = "FPGA, Low-power, NAS, challenge, computer vision, drone, knowledge distilling, model compression, scene text",

author = "Xiao Hu and Chang, {Ming Ching} and Yuwei Chen and Rahul Sridhar and Zhenyu Hu and Yunhe Xue and Zhenyu Wu and Pengcheng Pi and Jiayi Shen and Jianchao Tan and Xiangru Lian and Ji Liu and Zhangyang Wang and Liu, {Chia Hsiang} and Han, {Yu Shin} and Sung, {Yuan Yao} and Yi Lee and Wu, {Kai Chiang} and Guo, {Wei Xiang} and Rick Lee and Shengwen Liang and Zerun Wang and Guiguang Ding and Gang Zhang and Teng Xi and Yubei Chen and Han Cai and Ligeng Zhu and Zhekai Zhang and Song Han and Seonghwan Jeong and Youngmin Kwon and Tianzhe Wang and Jeffery Pan",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 3rd IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021 ; Conference date: 06-06-2021 Through 09-06-2021",

year = "2021",

month = jun,

day = "6",

doi = "10.1109/AICAS51828.2021.9458522",

language = "English",

series = "2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021",

address = "美國",

}

Hu, X, Chang, MC, Chen, Y, Sridhar, R, Hu, Z, Xue, Y, Wu, Z, Pi, P, Shen, J, Tan, J, Lian, X, Liu, J, Wang, Z, Liu, CH, Han, YS, Sung, YY, Lee, Y, Wu, KC, Guo, WX, Lee, R, Liang, S, Wang, Z, Ding, G, Zhang, G, Xi, T, Chen, Y, Cai, H, Zhu, L, Zhang, Z, Han, S, Jeong, S, Kwon, Y, Wang, T & Pan, J 2021, The 2020 Low-Power Computer Vision Challenge. 於 2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021., 9458522, 2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021, Institute of Electrical and Electronics Engineers Inc., 3rd IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021, Washington, 美國, 6/06/21. https://doi.org/10.1109/AICAS51828.2021.9458522

The 2020 Low-Power Computer Vision Challenge. / Hu, Xiao; Chang, Ming Ching; Chen, Yuwei 等.
2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021. Institute of Electrical and Electronics Engineers Inc., 2021. 9458522 (2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021).

研究成果: Conference contribution › 同行評審

TY - GEN

T1 - The 2020 Low-Power Computer Vision Challenge

AU - Hu, Xiao

AU - Chang, Ming Ching

AU - Chen, Yuwei

AU - Sridhar, Rahul

AU - Hu, Zhenyu

AU - Xue, Yunhe

AU - Wu, Zhenyu

AU - Pi, Pengcheng

AU - Shen, Jiayi

AU - Tan, Jianchao

AU - Lian, Xiangru

AU - Liu, Ji

AU - Wang, Zhangyang

AU - Liu, Chia Hsiang

AU - Han, Yu Shin

AU - Sung, Yuan Yao

AU - Lee, Yi

AU - Wu, Kai Chiang

AU - Guo, Wei Xiang

AU - Lee, Rick

AU - Liang, Shengwen

AU - Wang, Zerun

AU - Ding, Guiguang

AU - Zhang, Gang

AU - Xi, Teng

AU - Chen, Yubei

AU - Cai, Han

AU - Zhu, Ligeng

AU - Zhang, Zhekai

AU - Han, Song

AU - Jeong, Seonghwan

AU - Kwon, Youngmin

AU - Wang, Tianzhe

AU - Pan, Jeffery

PY - 2021/6/6

Y1 - 2021/6/6

N2 - AI computer vision has advanced significantly in recent years. IoT and edge computing devices such as mobile phones have become the primary computing platform for many end users. Mobile devices such as robots and drones that rely on batteries demand for energy efficient computation. Since 2015, the IEEE Annual International Low-Power Computer Vision Challenge (LPCVC) was held to identify energy-efficient AI and computer vision solutions. The 2020 LPCVC includes three challenge tracks: (1) PyTorch UAV Video Track, (2) FPGA Image Track, and (3) On-device Visual Intelligence Competition (OVIC) Tenforflow Track. This paper summarizes the 2020 winning solutions from the three tracks of LPCVC competitions. Methods and future directions for energy-efficient AI and computer vision research are discussed.

AB - AI computer vision has advanced significantly in recent years. IoT and edge computing devices such as mobile phones have become the primary computing platform for many end users. Mobile devices such as robots and drones that rely on batteries demand for energy efficient computation. Since 2015, the IEEE Annual International Low-Power Computer Vision Challenge (LPCVC) was held to identify energy-efficient AI and computer vision solutions. The 2020 LPCVC includes three challenge tracks: (1) PyTorch UAV Video Track, (2) FPGA Image Track, and (3) On-device Visual Intelligence Competition (OVIC) Tenforflow Track. This paper summarizes the 2020 winning solutions from the three tracks of LPCVC competitions. Methods and future directions for energy-efficient AI and computer vision research are discussed.

KW - FPGA

KW - Low-power

KW - NAS

KW - challenge

KW - computer vision

KW - drone

KW - knowledge distilling

KW - model compression

KW - scene text

UR - http://www.scopus.com/inward/record.url?scp=85113310185&partnerID=8YFLogxK

U2 - 10.1109/AICAS51828.2021.9458522

DO - 10.1109/AICAS51828.2021.9458522

M3 - Conference contribution

AN - SCOPUS:85113310185

T3 - 2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021

BT - 2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 3rd IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021

Y2 - 6 June 2021 through 9 June 2021

ER -

Hu X, Chang MC, Chen Y, Sridhar R, Hu Z, Xue Y 等. The 2020 Low-Power Computer Vision Challenge. 於 2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021. Institute of Electrical and Electronics Engineers Inc. 2021. 9458522. (2021 IEEE 3rd International Conference on Artificial Intelligence Circuits and Systems, AICAS 2021). doi: 10.1109/AICAS51828.2021.9458522

The 2020 Low-Power Computer Vision Challenge

摘要

出版系列

Conference

存取文件

其它文件與鏈接

指紋

引用此