Synthetic data set-based three-dimensional multi-person posture estimation

Abstract

The invention provides synthetic data set-based three-dimensional multi-person posture estimation. The main contents include a human body posture hierarchical skeleton, posture inference and a networkframework. A process comprises the steps of firstly giving an RGB image, estimating three-dimensional postures of multiple persons in the image, and coding joint positions relative to reference joints; secondly decomposing a body into trunk, four limbs and head by utilizing a special reading scheme of shielding a robust position graph, establishing vectors, and reading out corresponding limb bodypostures; and finally inferring three-dimensional postures through a reading priority and a two-dimensional joint verification policy. A new annotated multi-person data set is created by using an existing single-person three-dimensional data synthesis method; two-dimensional and three-dimensional joint positions can be predicted without bounding box extraction; even for complex multi-person scenes and shielding situations, person postures can be effectively predicted; and realization of the multi-person posture estimation greatly facilitates estimation of the person postures in actual scenes.

Description

technical field [0001] The invention relates to the field of posture estimation, in particular to a three-dimensional multi-person posture estimation based on a synthetic data set. Background technique [0002] Human body detection and pose estimation are one of the important foundations and research hotspots for action recognition in the field of computer vision. Person pose estimation has high application value. For example, the person pose estimation system can detect the position of the person from the image, and then analyze some additional information of the person (pose, age, gender, expression, identity and behavior, etc.) . By identifying and estimating the posture of the person, the workload of the staff who analyzes the posture of the person in the video can be reduced, and the cost of time, financial resources and material resources can be reduced. [0003] Human body posture estimation can be applied in the field of intelligent video surveillance. By analyzing...

AI Technical Summary

Problems solved by technology

However, traditional character pose estimation is difficult to deal with when the character's body is self-occluded, mutual occlusion between multiple characters, and other objects occlude the character. Among them, the pose estimation of multiple characters does not have a variety of data sets that can be used for training. making it difficult for the technology to achieve an effective breakthrough

Images