Rethinking Deconvolution for 2D Human Pose Estimation Light yet Accurate Model for Real-time Edge Computing

11/08/2021
by   Masayuki Yamazaki, et al.
0

In this study, we present a pragmatic lightweight pose estimation model. Our model can achieve real-time predictions using low-power embedded devices. This system was found to be very accurate and achieved a 94.5 HRNet 256x192 using a computational cost of only 3.8 model adopts an encoder-decoder architecture and is carefully downsized to improve its efficiency. We especially focused on optimizing the deconvolution layers and observed that the channel reduction of the deconvolution layers contributes significantly to reducing computational resource consumption without degrading the accuracy of this system. We also incorporated recent model agnostic techniques such as DarkPose and distillation training to maximize the efficiency of our model. Furthermore, we applied model quantization to exploit multi/mixed precision features. Our FP16'ed model (COCO AP 70.0) operates at  60-fps on NVIDIA Jetson AGX Xavier and  200 fps on NVIDIA Quadro RTX6000.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset