Image Processing Apparatus for Human Pose Template Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques face challenges in efficiently registering images with human bodies in desired poses and movements that differ from those indicated by preregistered template images, leading to reduced workability in template image registration.
Innovation Solution
An image processing apparatus comprising a skeleton structure detection unit, a similarity degree computation unit, a determination unit, and an output unit, which detects keypoints of human bodies, computes similarity between detected poses/movements and template images, determines areas with low similarity, and outputs these areas as candidates for additional template image registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a template image with a desired pose and movement is preregistered, then a human body with that pose and movement can be detected from images, but there is room for improvement in workability when registering additional template images with different poses and movements
Solution Approach 1:
The system automatically identifies candidate regions for new template images by detecting human bodies in images, computing their poses and movements, and selecting those with low similarity to existing templates. This self-service mechanism eliminates the need for manual selection of template candidates, significantly improving the ease of operation when registering additional templates with different poses and movements.
Solution Approach 2:
The system uses similarity degree computation based on pose and movement parameters to automatically determine which regions are suitable for new template registration. By changing the selection criterion from manual choice to automated similarity-based selection, the system improves workability while managing the complexity of handling multiple diverse templates.
2Productivity
If manual selection of template images is used, then templates can be registered, but the process becomes time-consuming and less efficient
Solution Approach 1:
The system performs preliminary detection of human bodies and computation of their poses and movements before template registration. By pre-identifying candidate regions and calculating their similarity to existing templates, the system prepares the necessary information in advance, enabling rapid and efficient template registration without manual intervention, thus improving productivity and reducing time loss.
Solution Approach 2:
The system provides feedback by outputting information about determined candidate regions for new template images. This feedback mechanism allows operators to quickly review and confirm template candidates without manual searching, significantly accelerating the template registration process and improving overall efficiency while minimizing time investment.
Data Source
AI summary
The present invention of an image processing apparatus (10) detects a keypoint of a human body included in an image; computes a degree of similarity between a pose or a movement of a human body detected from the image and a pose or a movement of a human body indicated by a preregistered template image; determines a place in the image where a human body also with the degree of similarity to a pose or a movement of a human body indicated by any template image to be less than a first threshold value is captured; and outputs information indicating the determined place or a partial image acquired by cutting the determined place out of the image.