Animation Model Production Using Markerless Image Subsets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer-based animation techniques face challenges in efficiently capturing and reproducing nuanced facial expressions and body movements, particularly in varying lighting conditions and across multiple sessions, due to the need for extensive marker application and complex image processing.
Innovation Solution
A system and method for selecting a subset of images from a set of captured images to produce a model, which includes representing surface features and estimating additional representations to enhance the model, using linear transformations like principle component analysis for efficient storage and reuse, allowing for the combination of images from different sessions and actors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If markers are applied to the actor's face to track facial expressions, then measurement precision is improved, but device complexity and ease of manufacture deteriorate due to extensive marker application requirements
Solution Approach 1:
The patent extracts and removes the markers from the system entirely, replacing them with markerless facial feature detection. The system now detects facial expressions by analyzing natural facial features and geometries directly from images, eliminating the need for physical marker application while maintaining measurement precision through advanced image processing algorithms
Solution Approach 2:
The patent replaces the mechanical marker application system with an automated computational system. Instead of physically applying markers to the actor's face, the system uses computer vision algorithms to detect and track facial features automatically, substituting mechanical processes with computational ones
2Manufacturing precision
If multiple cameras with different viewing perspectives are used to capture images, then manufacturing precision of the 3D model is improved, but device complexity worsens due to synchronization requirements
Solution Approach 1:
The patent makes the image capture system multi-functional by using a single camera that can capture images from multiple perspectives or positions. The camera system is designed to move or be repositioned to capture the actor from different angles, eliminating the need for multiple synchronized cameras while maintaining 3D modeling accuracy
Solution Approach 2:
The patent segments the image capture process into multiple sequential captures from different positions rather than requiring simultaneous captures from multiple cameras. The system captures images one at a time from different perspectives, then combines them computationally to create the 3D model, avoiding synchronization complexity
3Reliability
If a large set of captured images is used to produce a model, then reliability of the model is improved, but productivity deteriorates due to extensive processing time
Solution Approach 1:
The patent extracts and removes unnecessary images from the dataset, selecting only the most informative and representative images for model production. The system identifies and eliminates redundant or low-quality images, reducing the total number of images that need to be processed while maintaining model reliability through careful selection of high-value data
Solution Approach 2:
The patent uses a partial approach by processing only a subset of the captured images rather than all images. The system selectively processes a smaller number of key images that contain the most critical information for model creation, achieving sufficient model accuracy without the computational burden of processing the entire image set
Data Source
AI summary
A computer-implemented method includes selecting a subset of images from a set of captured images. A surface feature of one object is represented in the content of each selected subset image. The method also includes decomposing the surface feature content of each selected image to produce a model of representations of the object.