Towards Fast and Accurate Multi-Person Pose Estimation on Mobile Devices
This work addresses the need for efficient pose estimation on mobile platforms for applications like autonomous driving and behavior recognition, representing an incremental improvement.
The paper tackled the problem of high computational cost and latency in multi-person pose estimation on mobile devices by proposing an architecture optimization and weight pruning framework, achieving up to 2.51x faster inference speed with higher accuracy compared to existing lightweight methods.
The rapid development of autonomous driving, abnormal behavior detection, and behavior recognition makes an increasing demand for multi-person pose estimation-based applications, especially on mobile platforms. However, to achieve high accuracy, state-of-the-art methods tend to have a large model size and complex post-processing algorithm, which costs intense computation and long end-to-end latency. To solve this problem, we propose an architecture optimization and weight pruning framework to accelerate inference of multi-person pose estimation on mobile devices. With our optimization framework, we achieve up to 2.51x faster model inference speed with higher accuracy compared to representative lightweight multi-person pose estimator.