Hand Gesture Recognition via HOG Descriptors and Background Subtraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for recognizing hand gestures in complex backgrounds or with overlapping skin colors struggle with accuracy and efficiency, requiring extensive data sets and being vulnerable to lighting variations and noise.

Innovation Solution

The method involves comparing background and target images to detect candidate regions, generating HOG descriptors, and measuring resemblance values to identify hand gestures, using a HOS map to robustly recognize hand movements even in noisy conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If skin color detection is used to identify hand gestures, then the method can detect skin-colored regions, but it cannot distinguish between hand regions and other skin color regions in the background

Engineering Contradiction:
Improvehand detection accuracyVSAvoidfalse detection rate
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent segments the hand detection process into multiple independent analysis stages: skin color region detection, edge detection, contour analysis, and gesture pattern recognition. Each stage processes specific features separately and combines results to achieve accurate hand gesture identification while filtering out background skin-colored regions.

Inventive Principle:
Principle #1Segmentation

2Reliability

If Gaussian mixture model learning is conducted using extensive data sets, then a reliable model can be generated, but much time and effort are required for data preparation and model training

Engineering Contradiction:
Improvemodel reliabilityVSAvoiddata preparation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts essential gesture recognition features directly from image data without requiring extensive training datasets. It uses edge detection, contour extraction, and geometric feature analysis to identify hand gestures, eliminating the time-consuming model training process while maintaining recognition accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

3Device complexity

If traditional hand detection methods are used, then the system can operate with simple processing, but it is vulnerable to indoor lighting variations and camera noise

Engineering Contradiction:
Improveprocessing complexityVSAvoiddetection stability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent applies different processing techniques to different regions and features: skin color detection for region identification, edge detection for boundary definition, and contour analysis for shape characterization. This multi-faceted approach ensures robust hand gesture recognition under varying lighting conditions and camera noise while maintaining reasonable processing complexity.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS8548196B2Method and interface of recognizing user's dynamic organ gesture and elec tric-using apparatus using the interface
Publication Date: 2013.10.01 LG DISPLAY CO LTD
  • US8548196B2 patent drawing
  • US8548196B2 patent drawing
  • US8548196B2 patent drawing

AI summary

A method of recognizing a user's dynamic organ for use in an electric-using apparatus includes comparing a background image and a target image, which are inputted through an imaging element, to detect a candidate region including portions of the target image that are different between the background image and the target image; scanning the candidate region using a window; generating a HOG (histograms of oriented gradients) descriptor of a region of the target image that is scanned when it is judged that the scanned region includes a dynamic organ; measuring a resemblance value between the HOG descriptor of the scanned region and a HOG descriptor of a query template for a gesture of the dynamic organ; and judging that the scanned region includes the gesture of the dynamic organ when the resemblance value meets a predetermined condition.