3D Joint Gesture Recognition Using Static-Dynamic Model Fusion

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional hand gesture recognition technologies face high computational burdens, sensitivity to illumination and background, and inefficiencies in recognizing both static and dynamic gestures, often relying on large models that increase complexity and execution time.

Innovation Solution

A system utilizing two models for simultaneous static and dynamic gesture recognition, employing Gaussian error linear unit (GELU) activation and depth-wise separable convolutions, with a finite-state machine for static gesture recognition, and a fusion module to combine results, reducing computational complexity and improving performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a single large model is used to recognize both static and dynamic gestures, then gesture recognition capability is comprehensive, but computational complexity increases and execution speed decreases

Engineering Contradiction:

Improvegesture recognition capabilityVSAvoidcomputational complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent divides the gesture recognition system into separate static gesture recognition and dynamic gesture recognition modules. Each module is optimized for its specific function, with the static module using a smaller model and the dynamic module using a larger model, thereby reducing overall computational complexity while maintaining comprehensive recognition capability

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent creates a unified gesture recognition system that can handle both static and dynamic gestures through modular architecture. The system provides universal functionality by integrating multiple specialized modules that work together, allowing comprehensive gesture recognition without requiring a single oversized model

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Measurement precision

If multiple large models are used for gesture recognition, then recognition accuracy improves, but computational complexity and execution time increase

Engineering Contradiction:

Improverecognition accuracyVSAvoidexecution time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the recognition tasks into static and dynamic components, allowing each to be processed by optimized specialized models. This segmentation enables parallel processing and reduces the computational burden on each individual model, thereby decreasing execution time while maintaining high accuracy

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different levels of model complexity appropriate to each gesture type. Static gestures use a smaller, faster model while dynamic gestures use a more comprehensive model, avoiding the excessive computational action of applying a large model to all gesture types and thereby reducing overall execution time

Inventive Principle:

Principle #16Partial or excessive action

3Ease of operation

If conventional HGR technologies are used, then gesture recognition is achieved, but the system is sensitive to illumination and background conditions

Engineering Contradiction:

Improvegesture recognitionVSAvoidsensitivity to illumination and background

Core Design Contradiction:

Ease of operationVSObject-affected harmful factors

Solution Approach 1:

The patent extracts the gesture recognition process from dependence on illumination and background conditions by using skeleton-based representation. This extraction removes the harmful sensitivity to environmental factors while preserving the core gesture recognition functionality

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent replaces appearance-based recognition (which is sensitive to lighting) with skeleton-based recognition. This substitution uses structural joint information instead of visual appearance, eliminating sensitivity to illumination and background while maintaining gesture recognition capability

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

4Ease of operation

If appearance-based HGR is used, then gesture recognition is performed, but pre-processing requirements and computational burden increase

Engineering Contradiction:

Improvegesture recognitionVSAvoidpre-processing requirements

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent replaces appearance-based recognition with skeleton-based recognition, substituting complex image processing requirements with simpler skeletal joint data processing. This reduces pre-processing requirements and computational burden while maintaining gesture recognition functionality

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20250209854A1System and method for static and dynamic real-time gesture recognition

Publication Date: 2025.06.26 SAMSUNG ELECTRONICS CO LTD

AI summary

A system and a method are disclosed for performing gesture recognition. A method includes receiving frames of 3D physical body joints; performing a dynamic gesture recognition operation on a window of the frames; performing a static gesture recognition operation on an individual frame among the frames; applying a result of the static gesture recognition operation to a finite-state machine; fusing results of the dynamic gesture recognition operation and the finite-state machine; and generating a final recognized gesture based on the fusing.