Gesture detection and recognition method and system

• US 10,318,800 B2
• Filed: 06/13/2016
• Issued: 06/11/2019
• Est. Priority Date: 06/30/2015
• Status: Active Grant

• Current Assignee: Yutou Technology (Hangzhou) Co., Ltd.
• Original Assignee: Yutou Technology (Hangzhou) Co., Ltd.
• Inventors: Zhang, Hongxin
• Primary Examiner(s): Johns, Andrew W
• Application Number:
  US15/739,274
• Publication Number:
  US 20180293433A1
• Time in Patent Office: 1,093 Days
• CPC Class Codes: G06F 18/2431 Multiple classes
• CPC Class Codes: G06F 3/01 Input arrangements or combined input and output arrangements for interaction between user and computer G06F3/16 takes precedence
• CPC Class Codes: G06F 3/017 Gesture based interaction, e.g. based on a set of recognized hand gestures interaction based on gestures traced on a digitiser G06F3/04883
• CPC Class Codes: G06F 3/0304 Detection arrangements using opto-electronic means constructional details of pointing devices not related to the detection arrangement using opto-electronic means G06F3/033; optical digitisers G06F3/042
• CPC Class Codes: G06V 40/28 Recognition of hand or arm movements, e.g. recognition of deaf sign language static hand signs G06V40/113

Abstract

Disclosed are a gesture detection and recognition method and system. The gesture detection and recognition method comprises: acquiring and storing an image; adopting a plurality of pre-set classifiers for deleting different gestures to detect each frame of the image according to a pre-set sequence in a frame alternating manner to acquire a gesture target; establishing a skin color model based on the pixel distribution of the region of the gesture target; and acquiring gesture frequencies of the gesture target in a previous state and a later state according to the skin color model, and matching the gesture frequencies with a pre-set gesture state, so that a gesture change state is obtained and output. The gesture detection and recognition method and extract a skin color under specific scene, and can gradually eliminate the influence generated by acute illumination changes, thus realizing the purpose of extracting a gesture change state.

Claims

1. A gesture detection and recognition method, wherein:

the method comprises the steps of;

A1. collecting images and storing the images;

A2. using preset multiple classifiers used for detecting different gestures to detect each frame of the images in a way of alternating every two frames in preset order, to acquire gesture targets;

A3, establishing skin color models based on pixel distribution in target area of the gestures;

A4. acquiring gesture frequency of said gesture target in two consecutive states according to the skin color model, and match said gesture frequency with preset gesture state, to acquire gesture transition state and output.

2. The gesture detection and recognition method of claim 1, wherein:

said image is preprocessed before performing said step A2.

3. The gesture detection and recognition method of claim 1, wherein:

each of said classifiers performs multi-scale target detection of said image through a preset sliding window, to obtain said gesture target.

4. The gesture detection and recognition method of claim 3, wherein:

after said gesture target is obtained, said window is expanded by four times to detect said gesture target.

5. The gesture detection and recognition method of claim 1, wherein:

said classifiers employ cascade classifiers.

6. A gesture detection and recognition system, wherein:

the system comprises;

an acquisition unit for collecting images;

a storage unit connecting to said acquisition unit for storing said images;

a plurality of classifiers for detecting different gestures connecting to said storage unit for detecting each frame of said images in a way of alternating every two frames in preset order, to obtain gesture target;

a skin color modeling unit connecting to said storage unit for establishing skin color models based on pixel distribution in said target area of the gestures;

a decision making unit connecting to a plurality of said classifiers and said skin color modeling units respectively, acquiring said gesture frequency of said gesture targets in two consecutive states according to the skin color model, and match said gesture frequency with preset gesture state, to acquire gesture transition state and output.

7. The gesture detection and recognition system of claim 6, wherein:

said acquisition unit employs video camera.

8. The gesture detection and recognition system of claim 6, wherein:

said classifiers employ cascade classifiers.

9. The gesture detection and recognition system of claim 6, wherein:

said classifiers all perform multi-scale target detection of said image through a preset sliding window, to obtain said gesture target.

10. The gesture detection and recognition system of claim 9, wherein:

after said classifiers obtain said gesture target, said window will be expanded by four times, to detect said gesture target.

11. A gesture detection and recognition method, comprising the steps of:

providing an acquisition unit and a storage unit;

acquiring and storing a plurality of images;

providing a plurality of pre-set classifiers for detecting different gestures to detect each frame of said plurality of images according to a pre-set sequence in a frame alternating manner and acquiring a gesture target;

providing a skin color modeling unit based upon a pixel distribution of a region of said gesture target; and

providing a decision-making unit that acquires gesture frequencies of said gesture target in one of a previous state and a later state according to said skin color modeling unit.