Human Pose Estimation Using Parallel Architecture

Human posture recognition is extensively researched for identifying human activities and various applications. Even though it might be tough because of a number of things, like complex backgrounds, moving body positions, occlusions of the self and objects, poor image resolutions, and varying lighting, certain obstacles might occur. To address the restrictions and improve performance, a model for assessing human posture is developed that makes use of deep convolutional neural networks. The proposed method utilizes a bottom-up parsing technique to identify crucial locations in the human body. Moreover, it employs a non-parametric method to describe the vector field of key point associations, allowing for the grouping of anatomical key points for each individual. The accuracy of localized key points is further improved through multiple stages of enhanced prediction. The proposed method is trained and tested on the MPII Human Pose dataset, and it demonstrates superior performance in terms of accuracy compared to the latest state-of-the-art technique. Moreover, by including an occlusion network in the feature representation process, it efficiently discovers occluded key points, resulting in a mean average precision of 90.4%.