Date of Publication :8th August 2024

Abstract:Accurately detecting body keypoints in human pose estimation is essential for recognizing individuals' postures within an image. This pivotal process serves as a foundational requirement for various computer vision tasks, such as human action recognition, tracking, interactive interfaces, gaming, sign language interpretation, and video surveillance. After a concise overview, the classification as either single or multi-person pose estimation, determined by the quantity of individuals to be tracked, is presented. Subsequently, the narrative progresses by detailing the methodologies employed in human pose estimation. This is followed by an enumeration of applications and the challenges encountered within pose estimation. The focus then shifts towards a comprehensive examination of pivotal research, emphasizing their impact on human pose estimation. Each model's novelty, motivation, architectural design, operational procedures, practical applications, limitations, utilized datasets, and the evaluation metrics employed for assessing the model are succinctly discussed and analyzed.

Reference :