System and method for recognizing hand gesture

• US 10,013,070 B2
• Filed: 09/22/2016
• Issued: 07/03/2018
• Est. Priority Date: 03/29/2016
• Status: Active Grant

• Current Assignee: Korea Electronics Technology Institute (Government of South Korea)
• Original Assignee: Korea Electronics Technology Institute (Government of South Korea)
• Inventors: Ahn, Yang Keun, Park, Young Choong
• Primary Examiner(s): Sitta, Grant
• Application Number:
  US15/273,386
• Publication Number:
  US 20170285759A1
• Time in Patent Office: 649 Days
• Field of Search: 345156-173
• CPC Class Codes: G06F 3/005 Input arrangements through a video camera
• 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: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06V 40/11 Hand-related biometrics; Hand pose recognition

Abstract

Provided is a hand gesture recognition technique. A hand gesture recognition system according to one aspect of the present invention includes an image capturing unit that captures an image of a gesture recognition region to acquire a depth image; a hand region extractor that extracts a hand region from the depth image; and a hand gesture recognizer that analyzes the extracted hand region to recognize a hand gesture included in the depth image as a grasp gesture or a release gesture.

Claims

1. A system for recognizing a hand gesture using digital image processing, the system comprising:

an image capturing device configured to capture an image of a gesture recognition region to acquire a depth image;

a hand region extraction processor configured to extract a hand region from the depth image; and

a hand gesture recognition processor configured to analyze the extracted hand region to recognize a hand gesture included in the depth image as a grasp gesture or a release gesture, wherein the hand region recognition processor is further configured to primarily binarize the depth image, determine a hand region from the primarily binarized depth image, secondarily binarize the determined hand region, and finally extract the hand region,

wherein the hand region extraction processor is further configured to compare hand region candidates detected by applying a labeling algorithm to the primarily binarized depth image and a preset threshold, and determine a hand region candidate which is greater than or equal to the threshold as the hand region,

wherein the hand region recognition processor is further configured to adaptively binarize the extracted hand region to detect gesture recognition regions respectively corresponding to different fingers, and recognize the hand gesture as a release gesture or a grasp gesture on the basis of a number of the extracted gesture recognition regions, and

wherein the hand gesture recognition processor is further configured to recognize a hand gesture as a release gesture when the number of the gesture recognition regions is three or more, and wherein the hand gesture recognition processor is further configured to recognize a hand gesture as a grasp gesture when the number of the gesture recognition regions is two or less.

3. The system of claim 1, wherein whether the hand gesture is the grasp gesture or the release gesture is determined by recognizing an entire hand to increase a recognition rate of a hand gesture.

5. The system of claim 1, wherein the hand gesture recognition processor is further configured to recognize a hand gesture as a release gesture when the number of the gesture recognition regions is five.

2. A method of recognizing a hand gesture using digital image processing, the method comprising:

acquiring a depth image by capturing an image of a gesture recognition region;

analyzing the depth image and determining whether a hand region is extracted;

detecting gesture recognition regions respectively corresponding to different fingers by adaptively binarizing an extracted hand region when the hand region is extracted; and

recognizing a hand gesture included in the depth image as a release gesture or a grasp gesture on the basis of a number of the extracted gesture recognition regions, wherein the determining of whether a hand region is extracted includes primarily binarizing the depth image, determining a hand region from the primarily binarized depth image, secondarily binarizing the determined hand region, and finally extracting the hand region,

wherein the determining of whether a hand region is extracted includes comparing hand region candidates detected by applying a labeling algorithm to the primarily binarized depth image and a preset threshold, and determining a hand region candidate which is greater than or equal to the threshold as the hand region,

wherein the recognizing includes determining whether the number of the gesture recognition regions is greater than or equal to a preset value, and, when the number of the gesture recognition regions is greater than or equal to the preset value, recognizing the hand gesture as a release gesture, and when the number of the gesture recognition regions is less than the preset value, recognizing the hand gesture as a grasp gesture, and

wherein the recognizing further includes i) recognizing a hand gesture as a release gesture when the number of the gesture recognition regions is three or more and ii) recognizing a hand gesture as a grasp gesture when the number of the gesture recognition regions is two or less.

4. The method of claim 2, wherein the recognizing further includes recognizing an entire hand to determine whether a hand gesture is a grasp gesture or a release gesture to increase a recognition rate of a hand gesture.

6. The method of claim 2, wherein the recognizing further includes recognizing a hand gesture as a release gesture when the number of the gesture recognition regions is five.