Flexible Human Body Image Reshaping via Skeleton Key Point Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for reshaping human bodies in images lack flexibility, as they typically apply fixed transformations to the entire or partial human body, limiting the ability to perform region-specific and varying coordinate transformations.
Innovation Solution
A method and apparatus that determine a skeleton key point dataset from a target human body image to define a target region, allowing for coordinate transformation of pixel points within that region, using a reference skeleton key point dataset to establish transverse and longitudinal boundaries, and perform curve fitting to generate a coordinate transformation table for flexible reshaping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If fixed scale transformation is applied to the entire or partial human body, then the processing method is simple, but the flexibility and adaptability are poor
Solution Approach 1:
The patent divides the human body into multiple regions using skeleton key points (shoulder, elbow, wrist, hip, knee, ankle) as reference points. Each region is independently identified and processed, allowing different transformation operations on different body parts. This segmentation enables flexible, region-specific coordinate transformations while maintaining manageable processing complexity through systematic region identification and independent transformation application.
2Manufacturing precision
If region-specific coordinate transformation is performed, then the adaptability improves, but the processing complexity increases
Solution Approach 1:
The patent applies different transformation operations to different regions of the human body based on local requirements. Each region is processed with appropriate transformation parameters and methods suited to its specific characteristics, enabling precise control over the reshaping of individual body parts while maintaining overall coherence through the skeleton-based framework.
Solution Approach 2:
The patent transforms skeleton key point coordinates through coordinate transformation operations, using the transformed key points to define region boundaries and guide pixel point transformations. By changing the coordinate parameters of reference points and applying these transformations to corresponding image regions, the system achieves accurate regional reshaping with controllable complexity.
Data Source
AI summary
Techniques for processing an image are described herein. The disclosed techniques include a target human body image is acquired, and a skeleton key point data set is determined from the target human body image. Based on the skeleton key point data set, a target region in the target human body image is determined, and for each of pixel points comprised in the target region, performing coordinate transformation is performed on the pixel point, to generate a transformed coordinate of the pixel point. With the embodiment, different coordinate transformations may be performed on pixel points of different regions in the target human body image, thereby improving flexibility of performing the coordinate transformation on the pixel points.