Joint Optimization of Speed, Accuracy, and Energy for Embedded Image Recognition Systems
Duseok Kanga, DongHyun Kangb, Jintaek Kangc, Sungjoo Yood and Soonhoi Hae
Department of Computer Science and Engineering, Seoul National University, Seoul, Korea
akangds0829@snu.ac.kr
bkangdongh@snu.ac.kr
czealaton28@snu.ac.kr
dyeonbin@snu.ac.kr
esha@snu.ac.kr
ABSTRACT
This paper presents the image recognition system that won the first prize in the LPIRC (Low Power Image Recognition Challenge) in 2017. The goal of the challenge is to maximize the ratio between the accuracy and energy consumption within a time limit of 10 minutes for the processing of 20,000 images. Among three conflicting goals of accuracy, speed, and energy consumption, we considered the trade‐off between accuracy and speed first to select Nvidia Jetson TX2 as the hardware platform and Tiny YOLO as the image recognition algorithm. Next, we applied a series of software optimization techniques to improve throughput, such as pipelining, multithreading, Tucker decomposition, and 16‐bit quantization. Lastly, we explored the CPU and GPU frequencies to minimize the total energy consumption. As a result, we could achieve an accuracy of 0.24 mAP with energy consumption of 2.08Wh, which corresponds to the score of 0.11931, 2.7 times higher than the winner of LPIRC 2016.
Keywords: Embedded system, Energy consumption, Deep learning, Optimization, Neural networks.
