A human pose estimation algorithm based on the integration of improved convolutional neural networks and multi-level graph structure constrained model

摘要

This work introduces a novel convolutional feature for the task of human pose estimation, which is a framework of fusing a convolutional neural network into a multi-level graph structure model so as to improve the pose estimation results from body-part detection and human spatial structure. In the stage of part detection, the probability vectors corresponding to human body parts in the whole image are analyzed by a convolutional neural network. And then, a novel multi-level graph structure model is designed in this method, which contains the whole body, human body parts and joints, and realizes the coarse-to-fine establishment of a more constrained human spatiTal constraint model from levels of the structures, the edges, to the pixels. The obtained probability vectors are put into the multi-level graph structure model to compute the location coordinates for each joint, successfully achieving a framework of combining the deep learning network and the multi-level graph structure model. A large number of qualitative and quantitative experimental results show that compared with other state-of-art methods, the integration of the deep learning network and the multi-level pictorial structure model can improve the accuracy of human pose estimation to a greater extent.