Stereo Block Matching for Gesture Recognition Depth Maps

Solution Overview

Problem

Current gesture recognition technologies using depth extraction methods, particularly stereo block matching, are time-consuming and power-intensive, making them inefficient for real-time gesture recognition.

Innovation Solution

Implementing techniques such as auto-exposure, skin tone segmentation, determining a region of interest based on movement, and adaptive disparity range settings, along with feedback from higher-level applications, to selectively perform stereo block matching only on identified pixels, reducing computational load and power consumption while maintaining high performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If stereo block matching is performed on all pixels in the scene, then depth extraction accuracy is improved, but computational time and power consumption increase significantly

Engineering Contradiction:

Improvedepth extraction accuracyVSAvoidcomputational time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the image into multiple regions of interest (ROIs) based on gesture detection, and performs stereo block matching only within these segmented regions rather than on the entire image. This segmentation approach maintains depth extraction accuracy for gesture-related areas while significantly reducing computational time and power consumption by excluding background regions from processing.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different processing quality levels to different regions of the image. High-quality stereo block matching is applied locally to identified gesture regions, while other regions receive minimal or no processing. This local quality approach ensures accurate depth extraction where needed (gestures) while reducing overall computational burden.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If stereo block matching is performed on all pixels in the scene, then depth extraction accuracy is improved, but power consumption increases significantly

Engineering Contradiction:

Improvedepth extraction accuracyVSAvoidpower consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the image into multiple regions of interest (ROIs) based on gesture detection, and performs stereo block matching only within these segmented regions rather than on the entire image. This segmentation approach maintains depth extraction accuracy for gesture-related areas while significantly reducing computational time and power consumption by excluding background regions from processing.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different processing quality levels to different regions of the image. High-quality stereo block matching is applied locally to identified gesture regions, while other regions receive minimal or no processing. This local quality approach ensures accurate depth extraction where needed (gestures) while reducing overall computational burden.

Inventive Principle:

Principle #3Local quality

3Productivity

If adaptive disparity range settings are used based on feedback from higher-level applications, then processing efficiency is improved, but system complexity increases

Engineering Contradiction:

Improveprocessing efficiencyVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements a feedback mechanism where higher-level gesture recognition applications provide information about detected gestures back to the stereo processing module. This feedback enables adaptive adjustment of disparity range and ROI parameters, improving processing efficiency by focusing computational resources on relevant areas while maintaining manageable system complexity through structured feedback loops.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS9478037B2Techniques for efficient stereo block matching for gesture recognition

Publication Date: 2016.10.25 INTEL CORP

AI summary

Techniques to provide efficient stereo block matching may include receiving an object from a scene. Pixels in the scene may be identified based on the object. Stereo block matching may be performed for only the identified pixels in order to generate a depth map. Other embodiments are described and claimed.