Gesture recognition methods and systems

• US 8,417,026 B2
• Filed: 06/10/2010
• Issued: 04/09/2013
• Est. Priority Date: 12/30/2009
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Wu, Cheng-Feng, Tang, Cheng-Yuan, Chao, Shih-Pin
• Primary Examiner(s): COUSO, JOSE L
• Application Number:
  US12/813,464
• Publication Number:
  US 20110156999A1
• Time in Patent Office: 1,034 Days
• Field of Search: 382/103, 382/154, 382/165, 382/181, 382/190, 382/209, 345/156, 345/173, 345/175, 348 1407- 1408, 348/155, 348/169, 340/5.1, 340/12.22, 340/539.22, 701/1, 701/2, 701/408, 707/999.001, 715/716, 715/719, 715/773, 715/863
• US Class Current: 382/165
• 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 3/04883 for inputting data by handwriting, e.g. gesture or text
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

Gesture recognition methods and systems are provided. First, a plurality of gesture templates are provided, wherein each gesture template defines a first gesture characteristic and a corresponding specific gesture. Then, a plurality of images is obtained, and a multi-background model is generated accordingly. At least one object image is obtained according to the multi-background model, wherein the object image includes at least an object having a plurality of edges. The included angles of any two adjacent edges of the object image are gathered as statistics to obtain a second gesture characteristic corresponding to the object image. The second gesture characteristic of the object image is compared with the first gesture characteristic of each gesture template. The specific gesture corresponding to the first gesture characteristic is obtained, when the second gesture characteristic is similar to the first gesture characteristic.

Claims

1. A method of gesturing recognition executed by a computer, comprising:

providing a plurality of gesture templates, wherein each gesture template defines a first gesture characteristic and a specific gesture corresponding to the first gesture characteristic;

obtaining a plurality of images;

generating a multi-background model according to the images;

obtaining at least one object image according to the multi-background model, wherein the object image includes at least an object having a plurality of edges;

gathering statistics of the included angles of any two adjacent edges of the object image by the computer to obtain a second gesture characteristic corresponding to the object image;

comparing the second gesture characteristic of the object image with the first gesture characteristic of each gesture template; and

when the second gesture characteristic is similar to the first gesture characteristic, obtaining the specific gesture corresponding to the first gesture characteristic.

2. The method of claim 1, wherein each image has a plurality of pixels, each pixel has at least a color, and the generation of the multi-background model comprises the steps of:

for the respective pixel located at a location of the image, determining whether the appearance of a specific color corresponding to the pixels located at the same location of the respective images is greater than a predefined frequency;

determining that the pixel with the specific color as being representative of a part of a background when the appearance of the specific color corresponding to the pixels located at the same location of the respective images is greater than the predefined frequency; and

generating the multi-background model according to the respective pixels determined as being representative of the background and the corresponding colors thereof.

3. The method of claim 1, further comprising:

recognizing a specific object image from the at least one object image; and

calculating the corresponding second gesture characteristic for the specific object image.

4. The method of claim 3, wherein the specific object image comprises a hand image.

5. The method of claim 4, further comprising recognizing direction of a hand and location of fingertips corresponding to the hand image in the specific gesture.

6. The method of claim 1, wherein the second gesture characteristic has a similarity degree with each of the first gesture characteristics, and the highest similarity degree is greater than a predefined threshold.

7. The method of claim 1, wherein the second gesture characteristic is represented using a histogram, and the histogram records the statistical results of included angles of any two adjacent edges of the object image.

8. A gesture recognition system, comprising:

a storage unit storing a plurality of images and a plurality of gesture templates, wherein each gesture template defines a first gesture characteristic and a specific gesture corresponding to the first gesture characteristic; and

a processing unit generating a multi-background model according to the images, obtaining at least one object image according to the multi-background model, wherein the object image includes at least an object having a plurality of edges, gathering statistics of the included angles of any two adjacent edges of the object image to obtain a second gesture characteristic corresponding to the object image, comparing the second gesture characteristic of the object image with the first gesture characteristic of each gesture template, and obtaining the specific gesture corresponding to the first gesture characteristic when the second gesture characteristic is similar to the first gesture characteristic.

9. The system of claim 8, wherein each image has a plurality of pixels, each pixel has at least a color, and for the respective pixel located at a location of the image, the processing unit further determines whether the appearance of a specific color corresponding to the pixels located at the same location of the respective images is greater than a predefined frequency, determines that the pixel with the specific color as being representative of a part of a background when the appearance of the specific color corresponding to the pixels located at the same location of the respective images is greater than the predefined frequency, and generates the multi-background model according to the respective pixels determined as being representative of the background and the corresponding colors thereof.

10. The system of claim 8, wherein the processing unit further recognizes a specific object image from the at least one object image, and calculates the corresponding second gesture characteristic for the specific object image.

11. The system of claim 10, wherein the specific object image comprises a hand image.

12. The system of claim 11, wherein the processing unit further recognizes direction of a hand and location of fingertips corresponding to the hand image in the specific gesture.

13. The system of claim 8, wherein the second gesture characteristic has a similarity degree with each of the first gesture characteristics, and the highest similarity degree is greater than a predefined threshold.

14. The system of claim 8, wherein the second gesture characteristic is represented using a histogram, and the histogram records the statistical results of included angles of any two adjacent edges of the object image.

15. The system of claim 8, further comprising an image capture unit for capturing the images.

16. A machine-readable storage medium comprising a computer program, which, when executed, causes a device to perform a gesture recognition method, wherein the device has a plurality of images and a plurality of gesture templates, wherein each gesture template defines a first gesture characteristic and a specific gesture corresponding to the first gesture characteristic, and the method comprises:

generating a multi-background model according to the images;

obtaining at least one object image according to the multi-background model, wherein the object image includes at least an object having a plurality of edges;

gathering statistics of the included angles of any two adjacent edges of the object image to obtain a second gesture characteristic corresponding to the object image;

comparing the second gesture characteristic of the object image with the first gesture characteristic of each gesture template; and

obtaining the specific gesture corresponding to the first gesture characteristic when the second gesture characteristic is similar to the first gesture characteristic.