Flash Image Brightness Normalization via Region Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images captured in low-light environments using flash often suffer from unnatural brightness, red-eye effects, and irregular color balance due to the limited output power and irregular light emission of LED flashes in portable devices, leading to a decline in image quality.
Innovation Solution
An image processing method that adjusts the brightness of two images captured under different light conditions, one without a flash and one with a flash, by normalizing the brightness based on exposure times and estimating light intensity to separate regions, allowing for targeted brightness adjustments and color information overlay to achieve a final image with improved quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If flash is used to capture image in low-light environment, then brightness is improved, but color balance and image quality deteriorate due to irregular light emission
Solution Approach 1:
The patent segments the image processing into multiple stages: capturing two images under different light conditions (with and without flash), normalizing their brightness separately, estimating light intensity distribution, separating the image into multiple regions based on intensity, and applying different brightness adjustments to each region. This segmentation allows precise control over color balance in different areas while maintaining overall brightness.
Solution Approach 2:
The patent applies local quality by adjusting brightness differently for different regions of the image based on estimated light intensity. Regions with excessive brightness from flash are adjusted differently from regions with insufficient light, allowing each region to have optimal color balance and brightness characteristics suited to its specific lighting conditions.
2Measurement precision
If flash is used to capture image, then signal-to-noise ratio is improved, but brightness uniformity deteriorates due to foreground overexposure and background underexposure
Solution Approach 1:
The patent divides the image into multiple regions based on estimated light intensity distribution. This segmentation enables different brightness adjustment strategies for foreground regions (which may be overexposed) and background regions (which may be underexposed), thereby achieving uniform brightness across the entire image while preserving the high signal-to-noise ratio benefits of flash capture.
Solution Approach 2:
The patent changes the brightness parameter selectively for different regions of the image. By estimating light intensity and applying region-specific brightness adjustments, the system modifies the brightness parameter to compensate for uneven flash illumination, ensuring uniformity across foreground and background while maintaining the quality improvements from flash usage.
3Device complexity
If LED flash with low output power is used, then portability is improved, but image quality deteriorates due to insufficient and irregular light emission
Solution Approach 1:
The patent captures two images under different light conditions (with flash and without flash) and processes them separately through normalization and region-based adjustment. This segmentation compensates for the limitations of low-power LED flash by selectively using information from both images, thereby maintaining image quality despite the constraints of portable flash equipment.
Solution Approach 2:
The patent uses an intermediary processing approach where the final image is synthesized from two separate captures. By using the non-flash image as a complementary source and blending it with the flash image through region-based adjustments, the system overcomes the insufficient output power of the LED flash while maintaining portability.
Data Source
AI summary
An image processing method includes adjusting a brightness of each of a first image and a second image based on a first exposure time when the first image is captured and a second exposure time when the second image is captured, respectively, the first and second images being generated by capturing a same object under different light conditions, estimating an intensity of light, reaching the object when the second image is captured, based on the adjusted brightness of each of the first and second images, separating the second image into two or more regions according to the estimated intensity of light, determining a target brightness that a final result image is to have, adjusting the brightness of the second image, by different amounts for each of the separated regions, based on the target brightness and generate the final result image based on the adjusted brightness of the second image.