EHPE: Skeleton Cues-Based Gaussian Coordinate Encoding for Efficient Human Pose Estimation

Abstract

Human pose estimation (HPE) has many wide applications such as multimedia processing, behavior understanding and human-computer interaction. Most previous studies have encountered many constraints, such as restricted scenarios and RGB inputs. To mitigate constraints to estimating the human poses in general scenarios, we present an efficient human pose estimation model (i.e., EHPE) with joint direction cues and Gaussian coordinate encoding. Specifically, we propose an anisotropic Gaussian coordinate coding method to describe the skeleton direction cues among adjacent keypoints. To the best of our knowledge, this is the first time that the skeleton direction cues is introduced to the heatmap encoding in HPE task. Then, a multi-loss function is proposed to constrain the output to prevent the overfitting. The Kullback-Leibler divergence is introduced to measure the predication label and its ground truth one. The performance of EHPE is evaluated on two HPE datasets: MS COCO and MPII. Experimental results demonstrate that EHPE can obtain robust results, and it significantly outperforms existing state-of-the-art HPE methods. Lastly, we extend the experiments on infrared images captured by our research group. The experiments achieved the impressive results regardless of insufficient color and texture information.

Scoring method of English composition integrating deep learning in higher vocational colleges

“…The extended experiments were carried out on the collected infrared images. The results indicate that the experiment achieved good results when there was insufficient color and texture information 3 . Liu et al, designed an efficient deep matrix decomposition with retrospective feature learning for industrial recommendation systems to explain the characteristics of user reviews.…”

A Novel CNN Model for Classification of Chinese Historical Calligraphy Styles in Regular Script Font

“…Our literature review extends to other domains of machine learning that employ advanced CNN models for varied disciplinary objectives. For instance, Liu et al [ 21 ] developed TokenHPE (head pose estimation), a novel transformer-based model for head pose estimation. TokenHPE effectively utilizes orientation tokens to capture complex facial relationships, demonstrating state-of-the-art performance.…”

Research on the construction of a collaborative ability evaluation system for the joint graduation design of new engineering specialty groups based on digital technology

“…With the application of digital technology in various processes involving dental implants, a new type of dental implant treatment mode with high efficiency, comfort and accuracy has been formed. In this mode, the digital technology and equipment used for tooth defect treatment can obtain more accurate preoperative information, assist in conducting implant operations and impression repair, and complete personalized abutment design and production, thus generating significant value [ [12] , [13] , [14] ].…”