HOG Descriptor Calculation Using Bit-Shift Sector Classification

Solution Overview

Problem

Existing methods for calculating Histogram of Oriented Gradients (HOG) descriptors in image processing require multipliers and the arc tangent function, which are inefficient for hardware implementation, especially in devices with limited processing capability like hand-held cameras or smartphones.

Innovation Solution

A method that uses adders and comparators to determine the sector of gradient orientation without calculating exact angles, employing bit-shift operations and binary masks to map gradients into sectors, allowing for flexible bin combinations and efficient hardware implementation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multipliers and arc tangent functions are used to calculate HOG descriptors, then measurement precision of gradient orientation is improved, but device complexity and processing efficiency deteriorate

Engineering Contradiction:

Improvegradient orientation precisionVSAvoidhardware implementation complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transforms the continuous gradient orientation calculation into discrete sector classification by changing the parameter representation from continuous angles to discrete sector indices. This is achieved by comparing gradient components (Gx, Gy) against threshold values to determine which sector the gradient falls into, replacing the need for arc tangent calculations while maintaining sufficient precision for HOG descriptor generation.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mathematical computation system (multipliers and arc tangent functions) with a comparison-based logical system using bit-shift operations and binary masks. Instead of computing exact angles through trigonometric functions, the system uses sequential comparisons of gradient components against threshold values to classify gradients into sectors, significantly reducing hardware complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If multipliers and arc tangent functions are used to calculate HOG descriptors, then measurement precision of gradient orientation is improved, but processing speed deteriorates

Engineering Contradiction:

Improvegradient orientation precisionVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent transforms the continuous gradient orientation calculation into discrete sector classification by changing the parameter representation from continuous angles to discrete sector indices. This is achieved by comparing gradient components (Gx, Gy) against threshold values to determine which sector the gradient falls into, replacing the need for arc tangent calculations while maintaining sufficient precision for HOG descriptor generation.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mathematical computation system (multipliers and arc tangent functions) with a comparison-based logical system using bit-shift operations and binary masks. Instead of computing exact angles through trigonometric functions, the system uses sequential comparisons of gradient components against threshold values to classify gradients into sectors, significantly reducing hardware complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If standard HOG calculation methods are used, then measurement precision is improved, but ease of manufacture deteriorates

Engineering Contradiction:

Improvegradient orientation precisionVSAvoidhardware implementation ease

Core Design Contradiction:

Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent transforms the continuous gradient orientation calculation into discrete sector classification by changing the parameter representation from continuous angles to discrete sector indices. This is achieved by comparing gradient components (Gx, Gy) against threshold values to determine which sector the gradient falls into, replacing the need for arc tangent calculations while maintaining sufficient precision for HOG descriptor generation.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mathematical computation system (multipliers and arc tangent functions) with a comparison-based logical system using bit-shift operations and binary masks. Instead of computing exact angles through trigonometric functions, the system uses sequential comparisons of gradient components against threshold values to classify gradients into sectors, significantly reducing hardware complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS10839247B2Method for producing a histogram of oriented gradients

Publication Date: 2020.11.17 TOBII TECHNOLOGIES LTD

AI summary

A method for producing a histogram of oriented gradients (HOG) for at least a portion of an image comprises dividing the image portion into cells, each cell comprising a plurality of image pixels. Then, for each image pixel of a cell, obtaining a horizontal gradient component, gx, and a vertical gradient component, gy, based on differences in pixel values along at least a row of the image and a column of the image respectively including the pixel; and allocating a gradient to one of a plurality of sectors, where n is a sector index, each sector extending through a range of orientation angles and at least some of the sectors being divided from adjacent sectors according to the inequalities: b*16<a*(n+1) where 0≤n≤5, b=gy and a=gx; and b*16<a*(13−n) where 7≤n≤12, a=gx+gy and b=gx−gy. At least one sector is associated with a bin; and each instance of a pixel gradient of a cell associated with a bin is accumulated to provide a HOG for the cell.