Light-Weight RetinaNet for Object Detection on Edge Devices

TitleLight-Weight RetinaNet for Object Detection on Edge Devices
Publication TypeConference Proceedings
Year of Publication2020
AuthorsLi, YI, Dua, A, Ren, F
Conference NameThe 2020 IEEE World Forum on Internet of Things (WF-IoT'20)
Date Published04/2020
Conference LocationNew Orleans, Louisiana
Keywords (or New Research Field)psclab

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.Sx FLOPs reduction point over the original RetinaNet, and gains 1.Sx 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. 

File Attachment: