Method for analyzing object motion in multi frames

• US 8,478,002 B2
• Filed: 08/25/2011
• Issued: 07/02/2013
• Est. Priority Date: 04/01/2011
• Status: Active Grant

• Current Assignee: Altek Corporation
• Original Assignee: Altek Corporation
• Inventors: Peng, Shih-Yuan, Wu, Chung-Ta
• Primary Examiner(s): GORADIA, SHEFALI DINESH
• Application Number:
  US13/217,271
• Publication Number:
  US 20120250945A1
• Time in Patent Office: 677 Days
• Field of Search: 382/100, 382/107
• US Class Current: 382/107
• CPC Class Codes: G06T 7/223 using block-matching

Abstract

A method for analyzing object motion in multi frames adapted to an image capturing device is provided. Firstly, a plurality set of first sum of absolute difference (SAD) are obtained according to noise of the image capturing device under a plurality of light settings. Next, two frames are captured under a picturing light setting of the light settings. Then, a plurality of second SAD between the two frames are calculated. Afterwards, a plurality of object block within vein tracking of the object are found. Next, a local motion vector of each object block is respectively calculated according to the second SAD. Then, a first reliability of each object block is respectively calculated according to the second SAD and the set of the first SAD corresponding to the picturing light setting. Afterwards, the local motion vectors are estimated according to the first reliability to obtain a global motion vector.

Claims

1. A method for analyzing object motion in multi frames adapted to an image capturing device, comprising:

obtaining a plurality set of first sum of absolute difference according to a noise of the image capturing device under a plurality of light settings, wherein each of the light settings respectively corresponds to a set of first sum of absolute difference;

capturing a first frame and a second frame via the image capturing device under a picturing light setting of the light settings;

calculating a plurality of second sum of absolute difference between the first frame and the second frame;

finding a plurality of object block within a vein tracking of the object in the first frame and the second frame;

calculating a local motion vector of each object block respectively according to the second sum of absolute difference;

calculating a first reliability of each object block respectively according to the second sum of absolute difference and the set of first sum of absolute difference corresponding to the picturing light setting; and

estimating the local motion vectors according to the first reliability to obtain a global motion vector.

2. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 1, wherein the step of obtaining the sets of first sum of absolute difference according to the noise of the image capturing device under the light settings comprises:

capturing a plurality of images to a light box via the image capturing device under the light settings respectively; and

calculating and recording the first sum of absolute difference of the images in different locations.

3. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 2, wherein the step of calculating and recording the first sum of absolute difference of the images in different locations comprises:

recording the first sum of absolute difference of four corner pixels of each block of the images so as to interpolate the first sum of absolute difference of pixels of each block via the first sum of absolute difference of the four corner pixels.

4. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 2, wherein the step of calculating and recording the first sum of absolute difference of the images in different locations comprises:

recording the first sum of absolute difference shared by the pixels in each block of the images.

5. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 2, wherein the light source settings comprises:

at least one of brightness of the light box, color and photo sensitivity of the image capturing device and limb darkening.

6. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 1, wherein the step of finding the object blocks comprises:

calculating a maximum sum of absolute difference and a minimum sum of absolute difference of each block between the first frame and the second frame; and

determining whether each block is among one of the object blocks or a background block according to the difference between the maximum sum of absolute difference and the minimum sum of absolute difference of each block.

7. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 6, wherein the step of estimating the local motion vectors according to the first reliability to obtain the global motion vector comprises:

calculating a second reliability of the object blocks respectively according to the difference between the maximum sum of absolute difference and the minimum sum of absolute difference of each block; and

estimating the local motion vectors according to the first reliability and the second reliability to obtain the global motion vector.

8. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 7, wherein the step of estimating the local motion vectors according to the first reliability and the second reliability to obtain the global motion vector comprises:

calculating a plurality of average reliability according to the first reliability and the second reliability; and

weighting the local motion vectors according to the average reliability to obtain the global motion vector.

9. The method for analyzing object motion in multi frames adapted to an image capturing device as claimed in claim 1, wherein the step of calculating the first reliability of each object block respectively according to the second sum of absolute difference and the set of first sum of absolute difference corresponding to the picturing light setting comprises:

calculating a standard deviation tolerance of the set of first sum of absolute difference;

calculating a ratio of the second sum of absolute difference to the set of first sum of absolute difference; and

comparing the ratio and the standard deviation tolerance to obtain the first reliability.