Multi-Exposure Image Synthesis Using Pixel Luminance Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images with uneven luminance distribution due to partial bright and dark areas result in poor quality synthesis when combined, affecting user experience.
Innovation Solution
Determine luminance weights for each pixel based on the image luminance value and pixel luminance value of multiple images captured at different exposures, and perform image synthesis processing to adjust and combine these images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple images with different luminance are captured at different exposures and synthesized, then the synthesized image can cover a wider luminance range, but the synthesized image quality deteriorates when the image luminance distribution of the plurality of to-be-synthesized images is uneven
Solution Approach 1:
The patent applies local quality by calculating luminance weights for each pixel based on its local luminance value and the overall image luminance characteristics. This allows different regions of the image to have different weighting factors, ensuring that pixels with appropriate luminance contribute more to the synthesized image while pixels with extreme luminance values (too bright or too dark) contribute less, thereby maintaining uniform quality across the synthesized image.
Solution Approach 2:
The patent changes the parameter of luminance weight dynamically based on the luminance distribution characteristics. By calculating the luminance weight of each pixel according to its luminance value and the overall image luminance statistics, the system adapts the synthesis process to the specific luminance distribution of each image, resolving the quality issue caused by uneven luminance distribution.
2Device complexity
If simple image synthesis is performed on multiple images, then the processing complexity is low, but the synthesized image quality is poor when luminance distribution is uneven
Solution Approach 1:
The patent introduces feedback by calculating the luminance weight of each pixel based on the luminance distribution characteristics of the input images. The system analyzes the luminance values, determines weighting factors, and uses these weights to guide the synthesis process. This feedback mechanism ensures that the synthesis quality is maintained without requiring complex processing, as the weights automatically adjust to compensate for uneven luminance distribution.
3Adaptability or versatility
If images with great luminance differences are synthesized, then the synthesized image can represent both bright and dark regions, but the luminance uniformity of the synthesized image deteriorates
Solution Approach 1:
The patent applies local quality by assigning different luminance weights to pixels based on their local luminance characteristics. Pixels in regions with appropriate luminance values receive higher weights, while pixels in regions with extreme luminance values receive lower weights. This local differentiation ensures that the synthesized image maintains uniform luminance distribution while still representing both bright and dark regions of the scene.
Data Source
AI summary
Provided are image processing method and apparatus, an electronic device, and a computer-readable storage medium. The image processing method includes: obtaining at least two to-be-synthesized images, in which the at least two to-be-synthesized images are images of an object captured at different exposures; determining an image luminance value of each of the at least two to-be-synthesized images; determining, for each of the at least two to-be-synthesized images, a luminance weight of each pixel of the to-be-synthesized image based on the image luminance value of the to-be-synthesized image and a luminance value of each pixel of the to-be-synthesized image; and obtaining a target image by performing image synthesis processing based on the luminance value of each pixel of the at least two to-be-synthesized images and the luminance weight of each pixel of the at least two to-be-synthesized images. The method can improve quality of the synthesized image and user experience.