Image Synthesis Using Block-Based Change Detection Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-shot synthesis in photography often fails due to changes in imaging conditions such as object movement, camera shake, or changes in illumination, resulting in anomalous patterns and deteriorated image quality.
Innovation Solution
A photographing apparatus and method that sets one image as a reference and others as comparative images, generates reference and comparative maps by dividing them into blocks, detects changes between these maps to create a change-detection map, and synthesizes images while replacing affected parts with the reference image based on this map to maintain image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multi-shot synthesis is performed by sequentially photographing multiple images, then image definition and picture quality are improved, but image quality deteriorates when imaging conditions change during sequential photography
Solution Approach 1:
The patent divides the image into multiple blocks and processes each block independently. By segmenting the image into smaller regions, the system can detect changes in each block separately and apply appropriate synthesis strategies, preventing anomalous patterns from spreading across the entire image when imaging conditions change.
Solution Approach 2:
The patent implements a feedback mechanism by comparing sequentially captured images to detect changes in imaging conditions. The change detection result feeds back into the synthesis process, allowing the system to adjust its synthesis strategy based on detected changes, thereby maintaining image quality stability even when conditions vary.
2Measurement precision
If pixel shifting method is used for multi-shot synthesis, then higher definition is achieved, but net patterns and false colors occur when objects move during photography
Solution Approach 1:
The patent performs preliminary change detection by comparing sequentially captured images before performing the final synthesis. By detecting object movement and illumination changes in advance, the system can identify blocks that should not be synthesized together, preventing net patterns and false colors from occurring in the final image.
Solution Approach 2:
The patent applies different synthesis strategies to different blocks based on local change detection results. Blocks with significant changes are handled differently from stable blocks, allowing the system to maintain high definition in stable regions while avoiding artifacts in regions with object movement or illumination changes.
Data Source
AI summary
A photographing apparatus includes an image sensor which images a plurality of images; an image processor configured to set one image as a reference image and each remaining image as a comparative image; an image processor configured to generate a reference map divided into blocks; an image processor configured to generate a comparative map divided into blocks; an image processor configured to generate a change-detection map indicating coinciding or non-coinciding blocks between the reference map and the comparative map, per each respective block, upon the reference map and the comparative map being compared; an image processor configured to obtain a synthesized image of the images; and an image processor configured to replace at least a part of the synthesized image with a corresponding part of the reference image based on the change-detection map.