Median Filter Pixel Grouping for Hardware Resource Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Median filtering in image processing requires a large number of mathematical operations, particularly comparisons, which can be inefficient and resource-intensive, especially when implemented using hardware description language (HDL) customized for specific devices.
Innovation Solution
The proposed solution reduces the number of comparisons needed to determine a median parameter value for a neighborhood of pixels by employing techniques such as sorting and rearranging parameter values, allowing for the exclusion of subsets from comparison, thereby conserving hardware resources like comparators and multiplexers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If median filtering is implemented using traditional methods with hardware description language (HDL), then noise reduction in images is achieved, but the number of mathematical operations and hardware resources required increases significantly
Solution Approach 1:
The patent divides the pixel neighborhood into multiple groups (e.g., first group, second group, third group) and processes each group separately to determine median values. This segmentation reduces the total number of comparisons needed by avoiding exhaustive comparisons across all pixels simultaneously, thereby reducing hardware resource requirements while maintaining noise reduction effectiveness.
Solution Approach 2:
The patent performs preliminary sorting and rearranging of parameter values within each group before determining medians. By pre-organizing the data structure through sorting operations, the system reduces the number of comparisons required in subsequent median determination steps,从而降低硬件资源消耗 while保持图像处理效果。
2Reliability
If a large number of comparisons are performed to determine median parameter values, then accurate noise reduction is achieved, but processing time and computational load increase
Solution Approach 1:
By segmenting the pixel neighborhood into multiple groups and processing them in parallel, the system reduces the sequential processing time. Each group can be processed independently, allowing for concurrent operations that reduce total processing time while maintaining accurate median determination through structured comparison approaches.
Solution Approach 2:
The preliminary sorting and rearranging of parameter values within each group establishes an organized data structure that facilitates faster median determination. This pre-processing step reduces the number of comparisons needed during the actual median calculation, thereby decreasing processing time while maintaining noise reduction accuracy.
Data Source
AI summary
A device determines multiple sets of parameter values associated with an array of pixels to be used to process an image. Each set of parameter values corresponds to a set of pixels, included in the array, that are aligned along a first dimension of the array. The device determines a set of median values corresponding to each of the multiple sets of parameter values. The device determines a median of medians value that is a median of the set of median values. The device excludes a portion of parameter values, included in the multiple sets of parameter values, from a comparison set used to determine an overall median value, and determines a first median parameter value of the comparison set. The device determines the overall median value based on comparing the first median parameter value and the median of medians value, and provides the overall median value.