Image Processing Method for Pseudo-Flat Region Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Histogram equalization in video images can deteriorate the quality of pseudo-flat regions due to over-compression, leading to irregular blocks and decreased image quality.
Innovation Solution
An image processing method that generates a decision factor distribution image by calculating standard deviations and dividing it into marked regions, where pixels with smaller standard deviations are treated differently than those with larger deviations, to prevent enhancement of pseudo-flat regions during histogram equalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If histogram equalization is performed on the entire image to enhance contrast, then the overall image clarity is improved, but the quality of pseudo-flat regions deteriorates due to enlarged gray scale differences
Solution Approach 1:
The image is segmented into different regions based on local standard deviation calculations. Regions with standard deviation below a threshold are identified as pseudo-flat regions and excluded from histogram equalization, while other regions undergo the full enhancement process. This segmentation allows selective application of contrast enhancement to preserve pseudo-flat region quality.
Solution Approach 2:
Different processing strategies are applied to different regions of the image. Pseudo-flat regions maintain their original gray scale values to preserve quality, while non-pseudo-flat regions receive histogram equalization for contrast enhancement. This local quality approach ensures that each region is processed according to its specific characteristics.
2Measurement precision
If histogram equalization is applied uniformly to all pixels, then the gray distribution becomes uniform across the entire image, but irregular blocks appear in pseudo-flat regions
Solution Approach 1:
The image is divided into pseudo-flat regions and non-pseudo-flat regions using standard deviation analysis. Histogram equalization is selectively applied only to non-pseudo-flat regions, preventing the formation of irregular block artifacts while still achieving good gray distribution uniformity in the enhanced regions.
Solution Approach 2:
A decision factor distribution image is introduced as an intermediary to control the histogram equalization process. This intermediary image, derived from standard deviation calculations, guides which pixels should undergo equalization and which should remain unchanged, thereby preventing artifact formation while maintaining gray distribution uniformity.
3Manufacturing precision
If the standard deviation threshold is set low to preserve more pseudo-flat regions, then more regions maintain original quality, but fewer regions benefit from contrast enhancement
Solution Approach 1:
The standard deviation threshold parameter can be adjusted to control the balance between preserving pseudo-flat regions and enhancing contrast in other regions. By changing this parameter, users can optimize the trade-off between region preservation and enhancement coverage based on specific image characteristics and quality requirements.
Data Source
AI summary
An image processing method and device, storage medium and computer device are provided. The method includes: generating an original gray scale image of an original image; performing a histogram equalization process on the original gray scale image to obtain an equalized gray scale image; generating decision factor distribution image, wherein the decision factor distribution image includes a first marked region including a region where pixels that are adjacent in position and have standard deviations smaller than set value in the original image are located, and second marked region; obtaining final gray scale image according to original gray scale image, equalized gray scale image and decision factor distribution image. Gray scale values of pixel corresponding to second marked region and first marked region in final gray scale image are respectively gray scale values of corresponding pixel in equalized gray scale image and original gray scale image; and restoring a processed image.