Virtual Viewpoint Image Re-Synthesis for Failure Region Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques struggle to correct failures in generated volumetric images, requiring manual and inefficient correction work, especially for moving images, and lack support for easy correction of virtual viewpoint images.
Innovation Solution
An information processing system and method that allows users to designate failure regions in virtual viewpoint images, adjust color synthesis weights, and re-synthesize images based on user corrections, enabling easy and efficient correction of failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual retouch correction is performed for failures in volumetric images, then correction can be achieved, but the correction work becomes complicated and requires skilled techniques
Solution Approach 1:
The patent creates a library of pre-stored correction cases that map failure types to corrected versions. When a failure is detected, the system copies the appropriate correction case from the library and applies it automatically, eliminating the need for skilled manual retouch work while maintaining effective correction.
Solution Approach 2:
The correction system performs self-service by automatically detecting failure types, selecting appropriate correction cases from the library, and applying corrections without requiring user skill or manual intervention. The system serves itself by using stored correction knowledge to fix failures autonomously.
2Reliability
If correction cases are accumulated manually, then correction capability improves, but the accumulation process is time-consuming and inefficient
Solution Approach 1:
The patent performs preliminary action by pre-storing correction cases in a library before they are needed. Correction cases are accumulated and stored in advance, so when failures occur, the corrections are immediately available from the library without requiring time-consuming on-demand accumulation.
Solution Approach 2:
The system copies correction cases from the pre-stored library when failures are detected, rather than creating new corrections each time. This copying approach eliminates repetitive manual correction work and accelerates the correction process significantly.
3Reliability
If conventional correction methods are applied to moving images, then individual frame correction is possible, but correction work becomes a heavy burden due to continuous time series processing
Solution Approach 1:
The correction library stores correction cases that can be universally applied across multiple frames and time series. A single correction case can correct failures in multiple continuous frames simultaneously, making the correction process efficient for moving images without requiring individual frame processing.
Solution Approach 2:
The system copies correction cases from the library and applies them across multiple frames in the time series. This copying approach allows efficient correction of continuous moving images by reusing the same correction knowledge across all affected frames, dramatically improving productivity.
4Reliability
If color synthesis weights are adjusted for failure correction, then failure correction is achieved, but the correction process requires identifying and adjusting multiple weight parameters
Solution Approach 1:
The patent stores complete correction examples in the library that include the correct color synthesis weight values. When a failure is detected, the system copies the entire correction case including all necessary weight parameters, eliminating the complexity of manually identifying and adjusting multiple parameters.
Solution Approach 2:
The correction library acts as an intermediary that stores pre-computed correction information. Instead of directly adjusting multiple complex parameters, the system uses the library as an intermediary source that provides ready-to-apply correction cases with all parameters already optimized.
Data Source
AI summary
There is provided an information processing system capable of easily correcting a failure occurring in a virtual viewpoint image (free viewpoint image, volumetric image), an operation method of the information processing system, and a program. A virtual viewpoint image is generated by synthesizing on the basis of a color synthesis weight set for each of multi-viewpoint images, an input of a failure region of the virtual viewpoint image is received, a correction input of a color synthesis weight of the failure region of each of the multi-viewpoint images used for synthesis is received, the color synthesis weight is corrected to a plurality of the color synthesis weights based on the correction input, a plurality of the virtual viewpoint images is re-synthesized as a correction candidate image on the basis of the plurality of color synthesis weights, the color synthesis weights applied in a correction candidate image selected from a plurality of the correction candidate images as the failure is regarded as being corrected are determined as correction information for correcting the failure. The present disclosure can be applied to a generation device of a volumetric image.