Image processing apparatus, image processing method, recording medium, and program

• US 8,774,458 B2
• Filed: 01/20/2012
• Issued: 07/08/2014
• Est. Priority Date: 03/25/2011
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Ogawa, Kaname
• Primary Examiner(s): Abdi, Amara
• Application Number:
  US13/354,852
• Publication Number:
  US 20120243737A1
• Time in Patent Office: 900 Days
• Field of Search: 382/103, 382/118, 382/190, 382/201, 348/51, 348/E13.075
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation
• CPC Class Codes: G06T 7/248 involving reference images or patches

Abstract

An image processing apparatus includes: a calculating unit that calculates an evaluation value, which is expressed as a sum of confidence degrees obtained by mixing, at a predetermined mixing ratio, a matching degree of a first feature quantity and a matching degree of a second feature quantity between a target image containing an object to be tracked and a comparison image which is an image of a comparison region compared to the target image of a first frame, when the mixing ratio is varied and obtaining the mixing ratio when the evaluation value is maximum; and a detecting unit that detects an image corresponding to the target image of a second frame based on the confidence degrees in which the mixing ratio is set when the evaluation value is the maximum.

Claims

1. An image processing apparatus comprising:

a calculating unit that calculates an evaluation value, which is expressed as a sum of a first sum of products of a first weight coefficient and a first confidence degrees associated with each of a set of positive regions containing a part of a target image of a first frame containing an object to be tracked, and

a second sum of products of a second weight coefficient and a second confidence degrees associated with each of a set of negative regions not containing the target image of the first frame, wherein the first confidence degrees and the second confidence degrees are confidence degrees obtained by mixing, at a predetermined mixing ratio, a matching degree of a first feature quantity and a matching degree of a second feature quantity between the target image of the first frame and a comparison image which is an image of a comparison region from a set of comparison regions, wherein the mixing ratio is varied until the evaluation value is maximum, wherein the set of comparison regions is the set of positive regions when calculating the first confidence degrees and the set of comparison regions is the set of negative regions when calculating the second confidence degrees, wherein the first weight coefficient is a value obtained by dividing a constant by a number of positive regions in the set of positive regions and the second weight coefficient is a value obtained by dividing the constant by a number of negative regions in the set of negative regions; and

a detecting unit that detects a target image of a second frame using a third confidence degrees having a mixing ratio corresponding to the mixing ratio set when the evaluation value is the maximum.

2. The image processing apparatus according to claim 1, wherein the first frame corresponds to one of odd frames and the second frame corresponds to one of even frames.

3. The image processing apparatus according to claim 1, further comprising:

a computing unit that sets the set of comparison regions as scan regions on the second frame when calculating the third confidence degrees and calculates the third confidence degrees between the target image of the first frame and each scan image corresponding to each of the scan regions on the second frame.

4. The image processing apparatus according to claim 3, wherein the detecting unit detects a scan image, in which the third confidence degrees between the target image of the first frame and the scan image is the maximum, as the target image of the second frame.

5. The image processing apparatus according to claim 1, wherein the calculating unit sets an image in a reference region on the first frame as the target image of the first frame.

6. An image processing apparatus comprising:

a calculating unit that calculates an evaluation value, which is expressed as a sum of confidence degrees obtained by mixing, at a predetermined mixing ratio, a matching degree of a first feature quantity and a matching degree of a second feature quantity between a target image of a first frame containing an object to be tracked and a comparison image which is an image of a comparison region compared to the target image of the first frame, when the mixing ratio is varied and obtains the mixing ratio when the evaluation value is maximum;

a detecting unit that detects an image corresponding to a target image of a second frame based on the confidence degrees in which the mixing ratio is set when the evaluation value is the maximum, wherein the first and second frames are one of an odd frame and an even frame and the other thereof; and

a computing unit that sets a scan image in a scan region on the second frame as the comparison image and calculates the confidence degrees between the target image of the first frame and the scan image, wherein the detecting unit detects the scan image, in which the confidence degrees between the target image of the first frame and the scan image are the maximum, as an image corresponding to the target image of the second frame, wherein the calculating unit sets an image in a reference region on the first frame as the target image of the first frame, sets a plurality of regions containing at least a part of the target image of the first frame in the reference region as positive regions, sets a plurality of regions containing no target image of the first frame in the reference region as negative regions, calculates first confidence degrees which are the confidence degrees between the target image of the first frame in the reference region and the images in the plurality of positive regions, calculates second confidence degrees which are the confidence degrees between the target image of the first frame in the reference region and the images in the plurality of negative regions, calculates a first sum of products which is a sum of products of the first confidence degrees and a first weight coefficient of the positive regions, calculates a second sum of products which is a sum of products of the second confidence degrees and a second weight coefficient of the negative regions, and calculates a sum of the first sum of products and the second sum of products as the evaluation value, wherein the first weight coefficient is a value obtained by dividing a constant by the number of positive regions and the second weight coefficient is a value obtained by dividing the constant by the number of negative regions.

7. The image processing apparatus according to claim 6,

wherein the calculating unit calculates the evaluation value for a third frame subsequent to the second frame by setting an image in a region on the third frame, which corresponds to coordinates of the target image of the second frame, as a new target image of the third frame and obtains the mixing ratio when the evaluation value is the maximum, and

wherein the detecting unit detects an image corresponding to the new target image of the third frame on a fourth frame subsequent to the third frame based on the confidence degrees in which the mixing ratio is set when the evaluation value corresponding to the third frame is the maximum.

8. The image processing apparatus according to claim 6, further comprising:

a display unit that displays a marker on a region corresponding to coordinates corresponding either to the target image of the first frame or the target image of the second frame.

9. The image processing apparatus according to claim 6, further comprising:

a driving unit that drives a camera so that an image corresponding to the target image of the first frame is located at a predetermined position on a screen.

10. The image processing apparatus according to claim 6, wherein the first feature quantity is luminance information and the second feature quantity is color information.

11. An image processing method comprising:

calculating an evaluation value, which is expressed as a sum of

a first sum of products of a first weight coefficient and a first confidence degrees associated with each of a set of positive regions containing a part of a target image of a first frame containing an object to be tracked, and

a second sum of products of a second weight coefficient and a second confidence degrees associated with each of a set of negative regions not containing the target image of the first frame, wherein the first confidence degrees and the second confidence degrees are confidence degrees obtained by mixing, at a predetermined mixing ratio, a matching degree of a first feature quantity and a matching degree of a second feature quantity between the target image of the first frame and a comparison image which is an image of a comparison region from a set of comparison regions, wherein the mixing ratio is varied until the evaluation value is maximum, wherein the set of comparison regions is the set of positive regions when calculating the first confidence degrees and the set of comparison regions is the set of negative regions when calculating the second confidence degrees, wherein the first weight coefficient is a value obtained by dividing a constant by a number of positive regions in the set of positive regions and the second weight coefficient is a value obtained by dividing the constant by a number of negative regions in the set of negative regions; and

detecting a target image of a second frame using a third confidence degrees having a mixing ratio corresponding to the mixing ratio set when the evaluation value is the maximum.

12. A non-transitory computer readable medium having stored there on a computer program for image processing, wherein the computer program when executed by a computer causes the computer to perform steps of:

calculating an evaluation value, which is expressed as a sum of

a first sum of products of a first weight coefficient and a first confidence degrees associated with each of a set of positive regions containing a part of a target image of a first frame containing an object to be tracked and

a second sum of products of a second weight coefficient and a second confidence degrees associated with each of a set of negative regions not containing the target image of the first frame, wherein the first confidence degrees and the second confidence degrees are confidence degrees obtained by mixing, at a predetermined mixing ratio, a matching degree of a first feature quantity and a matching degree of a second feature quantity between the target image of the first frame and a comparison image which is an image of a comparison region from a set of comparison regions, wherein the mixing ratio is varied until the evaluation value is maximum, wherein the set of comparison regions is the set of positive regions when calculating the first confidence degrees and the set of comparison regions is the set of negative regions when calculating the second confidence degrees, wherein the first weight coefficient is a value obtained by dividing a constant by a number of positive regions in the set of positive regions and the second weight coefficient is a value obtained by dividing the constant by a number of negative regions in the set of negative regions; and

detecting a target image of a second frame using a third confidence degrees having a mixing ratio corresponding to the mixing ratio set when the evaluation value is the maximum.