TY - GEN
T1 - Light-Weight RetinaNet for Object Detection on Edge Devices
AU - Li, Yixing
AU - Dua, Akshay
AU - Ren, Fengbo
N1 - Funding Information:
This work is supported by an NSF grant (IIS/CPS-1652038) and an unrestricted gift from Radius AI, Inc. The computing infrastructure used in this work is supported by an NFS grant (CNS-1629888). The Arria 10 GX FPGA Development Kits used for this research was donated by Intel FPGA University Program. We thank Dr. Aykut Dengi and Bobby Chowdary from Radius AI, Inc. for fruitful research discussions.
Publisher Copyright:
© 2020 IEEE.
PY - 2020/6
Y1 - 2020/6
N2 - This paper aims at reducing computation for Retinanet, an mAP-30-tier network, to facilitate its practical deployment on edge devices for providing IoT-based object detection services. We first validate RetinaNet has the best FLOP-mAP trade-off among all mAP-30-tier network. Then, we propose a light-weight RetinaNet structure with effective computation- accuracy trade-off by only reducing FLOPs in computationally intensive layers. Compared with the most common way of trading off computation with accuracy-input image scaling, the proposed solution shows a consistently better FLOPs-mAP trade-off curve. Light-weight RetinaNet achieves a 0.3% mAP improvement at 1.8x FLOPs reduction point over the original RetinaNet, and gains 1.8x more energy-efficiency on an Intel Arria 10 FPGA accelerator in the context of edge computing. The proposed method potentially can help a wide range of the object detection applications to move closer to a preferred corner for a better runtime and accuracy, while enjoys more energy-efficient inference at the edge.
AB - This paper aims at reducing computation for Retinanet, an mAP-30-tier network, to facilitate its practical deployment on edge devices for providing IoT-based object detection services. We first validate RetinaNet has the best FLOP-mAP trade-off among all mAP-30-tier network. Then, we propose a light-weight RetinaNet structure with effective computation- accuracy trade-off by only reducing FLOPs in computationally intensive layers. Compared with the most common way of trading off computation with accuracy-input image scaling, the proposed solution shows a consistently better FLOPs-mAP trade-off curve. Light-weight RetinaNet achieves a 0.3% mAP improvement at 1.8x FLOPs reduction point over the original RetinaNet, and gains 1.8x more energy-efficiency on an Intel Arria 10 FPGA accelerator in the context of edge computing. The proposed method potentially can help a wide range of the object detection applications to move closer to a preferred corner for a better runtime and accuracy, while enjoys more energy-efficient inference at the edge.
UR - http://www.scopus.com/inward/record.url?scp=85095590000&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85095590000&partnerID=8YFLogxK
U2 - 10.1109/WF-IoT48130.2020.9221150
DO - 10.1109/WF-IoT48130.2020.9221150
M3 - Conference contribution
AN - SCOPUS:85095590000
T3 - IEEE World Forum on Internet of Things, WF-IoT 2020 - Symposium Proceedings
BT - IEEE World Forum on Internet of Things, WF-IoT 2020 - Symposium Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 6th IEEE World Forum on Internet of Things, WF-IoT 2020
Y2 - 2 June 2020 through 16 June 2020
ER -