Deep Learning Gesture Recognition With Segment-Level Result Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing gesture recognition methods based on computer vision suffer from excessive processing time due to segmentation and tracking of gestures, leading to significant delays.

Innovation Solution

A gesture recognition method utilizing deep learning algorithms to process multiple images from a video segment without segmentation or tracking, combining optical flow and color information for improved accuracy and speed, and employing result combination across consecutive video segments to enhance recognition performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If segmentation and tracking based on specific image tracking algorithm is used to extract gesture shape, feature, and location information, then gesture recognition accuracy is improved, but processing time increases significantly leading to excessive delay

Engineering Contradiction:
Improvegesture recognition accuracyVSAvoidgesture recognition delay
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The video stream is divided into multiple video segments, and each video segment is further divided into multiple image frames. This segmentation allows the system to process smaller units of data independently, reducing the overall processing time while maintaining recognition accuracy through sequential analysis of segmented data

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary processing by pre-dividing the video stream into segments and extracting key features from each segment before final gesture recognition. This preliminary action prepares the data in advance, reducing the time required for the final recognition decision and thereby reducing overall delay

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3651055B1Gesture recognition method, apparatus, and device
Publication Date: 2026.02.25 HUAWEI TECH CO LTD
  • EP3651055B1 patent drawingFigure 1~2
  • EP3651055B1 patent drawingFigure 3
  • EP3651055B1 patent drawingFigure 4(a)~5

AI summary

This application provides a gesture recognition method, and relates to the field of man-machine interaction technologies. The method includes: extracting M images from a first video segment in a video stream; performing gesture recognition on the M images by using a deep learning algorithm, to obtain a gesture recognition result corresponding to the first video segment; and performing result combination on gesture recognition results of N consecutive video segments including the first video segment, to obtain a combined gesture recognition result. In the foregoing recognition process, a gesture in the video stream does not need to be segmented or tracked, but phase actions are recognized by using a deep learning algorithm with a relatively fast calculation speed, and then the phase actions are combined, so as to improve a gesture recognition speed, and reduce a gesture recognition delay.