Character Image Data Processing via Head and Face Component Replacement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high cost and labor-intensive process of modeling digital virtual characters require extensive data collection in professional studios, increasing production costs due to the need for multiple recordings and professional guidance.
Innovation Solution
A method and apparatus for processing character image data that involves acquiring videos of character images, replacing head and face images to combine movement-state and attribute images from different characters, resulting in a target head image with improved clarity and rich movement-state images, thereby reducing the need for extensive data collection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If extensive data collection is performed in professional studios with multiple recordings and professional guidance, then the quality and richness of character image data is improved, but the production cost and time increase significantly
Solution Approach 1:
The patent applies copying by extracting and reusing head images and movement-state images from different video sources. Instead of recording all data in professional studios, the system copies relevant image components from existing videos and combines them through image processing to generate virtual character images, significantly reducing the need for extensive new data collection.
Solution Approach 2:
The patent segments character image data into distinct components: head images, movement-state images, and body images. This segmentation allows independent processing and combination of these elements from different source videos, enabling the system to assemble high-quality virtual character images without requiring complete re-recording of all data in professional studios.
2Reliability
If extensive data collection is performed in professional studios with multiple recordings and professional guidance, then the quality and richness of character image data is improved, but the production cost increases
Solution Approach 1:
The system copies reusable image components from existing video materials rather than investing in expensive professional studio recording sessions. By extracting head images and movement-state images from available sources and combining them through computational methods, the patent eliminates the need for costly repeated recordings while maintaining data quality.
Solution Approach 2:
The patent creates a universal image processing framework that can combine elements from any video source, not requiring specialized studio recordings. The system processes and combines images from diverse sources to generate virtual character images, making the production process more flexible and cost-effective while maintaining consistent output quality.
3Quantity of substance
If head images and movement-state images are combined from different characters, then the richness of movement-state images is improved, but the clarity of the face may be reduced
Solution Approach 1:
The patent applies local quality by applying different processing strategies to different parts of the image. The face replacement unit specifically processes the face region with higher precision to maintain clarity, while the head image replacing unit handles the overall head composition including movement-state images. This localized quality control ensures that the face maintains high clarity even when combining elements from different sources.
Solution Approach 2:
The system segments the face region from the overall head image and processes them through different units. The face replacing unit extracts and replaces face images with higher precision processing, while the head image replacing unit handles the broader head composition. This segmentation allows the face clarity to be optimized independently while still benefiting from the rich movement-state images combined from different characters.
Data Source
AI summary
A method and apparatus for processing character image data are provided. A specific embodiment of the method can include: acquiring a video including character images of a first character as first data, and acquiring a video including character images of a second character as second data, the video including character images of the first character including a plurality of types of movement-state images; replacing a head image extracted from the first data with a head image extracted from the second data; replacing a face extracted from the first data with a face extracted from the second data; and replacing the face in the replaced head image with the replaced face to obtain a target head image.