Gradient Magnitude Histogram Texture Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing texture descriptors in image processing, such as HOG and LBP, are sensitive to image rotation and computationally expensive, especially when dealing with high-dimensional features like RLBP and Rotation Invariant HOG, which complicates texture determination in real-world applications where images are often captured from varying angles.
Innovation Solution
A computer-implemented system and method for image texture determination that generates a histogram of gradient magnitude (HGM), which is rotational invariant, reducing computational complexity and storage requirements by quantizing magnitude values into 16 bins, allowing for efficient comparison with predetermined histograms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rotation-invariant texture descriptors like RLBP and Rotation Invariant HOG are used, then rotation sensitivity is reduced, but computational complexity and storage requirements increase
Solution Approach 1:
The patent extracts only the magnitude information from gradient computations, discarding the orientation component. This is achieved by computing gradient magnitude as G(i,j) = sqrt(Gx^2 + Gy^2) and then quantizing it into discrete bins, creating a histogram that is inherently rotation-invariant while significantly reducing computational complexity compared to full HOG or RLBP descriptors.
Solution Approach 2:
The patent changes the parameter representation by quantizing continuous gradient magnitude values into discrete bins (e.g., 16 bins). This transformation converts the continuous magnitude range into discrete intervals, creating a histogram representation that is computationally efficient and rotation-invariant, as the histogram distribution remains unchanged under rotation.
2Measurement precision
If high-dimensional texture features like LBP and RLBP are used, then texture description accuracy is improved, but storage costs and computational time increase
Solution Approach 1:
The patent transforms the texture representation by creating histograms of gradient magnitudes with a fixed number of bins (e.g., 16 bins). This parameter transformation reduces the dimensionality from high-dimensional feature vectors (256 dimensions for LBP, 36 dimensions for RLBP) to a compact histogram representation, significantly reducing storage requirements while maintaining texture discrimination capability.
Solution Approach 2:
The patent uses a simplified gradient magnitude histogram representation that requires minimal storage compared to high-dimensional texture descriptors. The histogram with 16 bins creates a compact, efficient representation that can be stored and processed with much lower memory requirements, effectively replacing complex high-dimensional features with a simpler, more efficient structure.
Data Source
AI summary
This disclosure relates generally to image processing, and more particularly to image texture determination. In one embodiment a processor a memory coupled to the processor, wherein the processor coupled with a plurality of modules stored in the memory: At least one image having a plurality of pixels is accepted. Any noise is removed from the image to obtain at least one noise free image. The at least one noise free image is converted to at least one gray scale image. Horizontal and Vertical Gradient for plurality of pixels of the at least one gray scale image are computed. Gradient magnitude is calculated for the generated gradient. Histogram of the gradient magnitude is generated based on the gradient magnitude, and the plurality of generated histograms are compared with a plurality of predetermined histograms.