GESTURE IDENTIFICATION WITH NATURAL IMAGES

• US 20140105453A1
• Filed: 10/04/2013
• Published: 04/17/2014
• Est. Priority Date: 10/08/2012
• Status: Abandoned Application

• Current Assignee: PixArt Imaging Incorporated
• Original Assignee: PixArt Imaging Incorporated
• Inventors: YANG, Shu-Sian
• Application Number:
  US14/046,214
• Publication Number:
  US 20140105453A1
• Time in Patent Office: Days
• US Class Current: 382/103
• 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

A method for gesture identification with natural images includes generating a series of variant images by using each two or more successive ones of the natural images, extracting an image feature from each of the variant images, and comparing the varying pattern of the image feature with a gesture definition to identify a gesture. The method is inherently insensitive to indistinctness of images, and supports the motion estimation in axes X, Y, and Z without requiring the detected object to maintain a fixed gesture.

Claims

1. A method for gesture identification with natural images, comprising the steps of:

A.) generating a time-based series of images;

B.) from the series of the images, each time selecting two or more successive images to generate a variant image;

C.) extracting an image feature from the variant image; and

D.) comparing a varying pattern of the image feature with a gesture definition to identify a gesture.

2. The method of claim 1, wherein the step B comprises the step of subtracting between the images being successive in terms of time.

3. The method of claim 1, wherein the step C comprises the step of screening pixels of the variant image by a threshold for calculating the image feature.

4. The method of claim 1, wherein the step C comprises the step of calculating a barycenter, a standard deviation or a variance of the variant image as the image feature.

5. The method of claim 1, wherein the step D comprises the step of processing the image feature by a frequency analysis to determine the varying pattern.

6. A method for gesture identification with natural images, comprising the steps of:

A.) generating a time-based series of images;

B.) comparing two or more successive images to generate a differential image from the series of the images; and

C.) comparing the differential image with a gesture definition to identify a gesture.

7. The method of claim 6, wherein the step B comprises the step of subtracting between the images being successive in terms of time.

8. The method of claim 6, wherein the step C comprises the step of screening pixels of the variant image by a threshold for calculating an image feature.

9. The method of claim 6, wherein the gesture includes a hand left motion, a hand right motion or a hand rotation motion.

10. The method of claim 6, wherein the step C comprises the step of processing the differential image by a frequency analysis to determine a varying pattern.

11. The method of claim 10, wherein the step C comprises the step of comparing the varying pattern of the differential image with the gesture definition to identify the gesture.