Multi-Person Pose Estimation Using Group-Based Convolutional Neural Network Model

Human pose estimation has drawn extensive attention recently and there been significant progress on it due to the rising popularity of convolutional neural networks (CNN). However, existing state-of-the-art approaches suffer from occlusion, complicated backgrounds, substantial position fluctuations because disregarding human body form. parsing is a very pertinent activity that can provide crucial semantic data about bodily parts for estimation. To overcome aforesaid limitations, this paper introduces method using group-based network model. The proposed adopts bottom-up strategy yields features extract skeletal key points in body. Moreover, creates grouping anatomical individuals by utilizing non-parametric description point association vector field. Experimental results indicate provides superior performance than algorithms terms accuracy. In addition, optimizes its output detects occluded as well invisible incorporating feature representation. surpasses recent methods, achieving 93% mean average