Method and system for searching images with figures and recording medium storing metadata of image

• US 8,170,343 B2
• Filed: 09/02/2007
• Issued: 05/01/2012
• Est. Priority Date: 06/25/2007
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Corel Corporation (Corel Holdings LP)
• Inventors: Chien, Yi-Tsung
• Primary Examiner(s): Repko, Jason M
• Assistant Examiner(s): Thomas, Mia M
• Application Number:
  US11/849,294
• Publication Number:
  US 20080317353A1
• Time in Patent Office: 1,703 Days
• Field of Search: 382/117, 382/118, 340/5.53, 340/5.83
• US Class Current: 382/194
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A method and a system for searching images with figures and a recording medium storing metadata of image are provided. The searching method is divided into an image analysis stage and an image search stage. In the image analysis stage, figures between images are compared with each other and assigned with an identity respectively. A representative image of each identity is then evaluated from the image collection. In the image search stage, the representative images are displayed for user to select some of them as a searching criterion, so as to search and display the images matching the searching criterion in the image collection. Accordingly, the images required by user can be found through intelligent analysis of figures, intuitive definition of searching criterion, and simple comparison of identities so that both time and effort of organization for searching images with figures can be substantially saved.

Claims

1. A method for searching images with figures, suitable for searching a plurality of images related to at least one figure in an image collection, the method comprising:

an image analysis stage, comprising;

comparing figures between images of an image collection and respectively assigning an identity to each image of the image collection for each occurrence of an identified figure of the figures, the identity not derived mathematically from the pixel content of the image and relating to the identified figure; and

selecting and associating a representative image in the image collection for each of the identities assigned; and

an image search stage, comprising;

displaying the representative images;

selecting one or a plurality of the representative images as a searching criterion;

searching for images matching the searching criterion in the image collection, the matching performed in dependence upon at least the identity associated with each selected representative image; and

displaying the images matching the searching criterion.

2. The search method according to claim 1, wherein before comparing figures between images of an image collection and respectively assigning an identity to each of the figures, the search method further comprises:

normalizing one of the color, brightness, and contrast of the images of the image collection.

3. The search method according to claim 1, wherein the step of comparing figures between images of an image collection and respectively assigning an identity to each of the figures comprises:

locating a characteristic region of each of the figures in the images;

extracting a regional characteristic of each of the figures in each of the characteristic regions; and

comparing the regional characteristic of each of the figures between images, and assigning an identity to the figures having the same regional characteristic.

4. The search method according to claim 3, wherein the characteristic region comprises one of a face region, a dress region, a hair region, and a background region.

5. The search method according to claim 3, wherein the step of comparing figures between images of an image collection and respectively assigning an identity to each of the figures comprises:

collecting a plurality of image groups in which figures have been respectively assigned with identities;

extracting the regional characteristic of each of the identities in the characteristic regions of the image groups; and

comparing the regional characteristic of each of the identities between the plurality of image groups, and

reassigning those identities having the same regional characteristic with a unified identity.

6. The search method according to claim 1, wherein the step of respectively selecting the representative image for each of the identities in the image collection comprises:

respectively calculating a representative index of the identity in each of the images; and

selecting the image having the largest representative index as the representative image of the identity in the image collection.

7. The search method according to claim 6, wherein the representative index comprises one of a size of the characteristic region, a number of appearances, and a capture time.

8. The search method according to claim 1, wherein the method for assigning the identity comprises recording the identity into a metadata corresponding to the image.

9. The search method according to claim 8, wherein the metadata is stored in a header file of the image or an external file corresponding to the image.

10. The search method according to claim 1, wherein the step of displaying the representative images comprises:

sorting the representative images according to a relationship index of each of the identities in the image collection; and

sequentially displaying the representative images.

11. The search method according to claim 10, wherein the relationship index comprises one of an average size of the characteristic regions, a number of appearances, and a capture time.

12. The search method according to claim 1, wherein the step of selecting one or a plurality of the representative images as the searching criterion comprises:

translating the selected representative images into the corresponding identities; and

using the identities as the searching criterion.

13. The search method according to claim 1, wherein the step of displaying the images matching the searching criterion comprises:

sorting the images matching the searching criterion according to the number of the figures or the size of characteristic region; and

sequentially displaying the images matching the searching criterion.

14. The search method according to claim 1, wherein the method further comprises:

storing metadata of each image of the image collection in one or more non-transitory tangible computer-readable storage media, the metadata comprising;

a facial characteristic data recording a facial data of a figure in the image;

a regional characteristic data recording a regional data of a characteristic region other than the face region in the image; and

an identification data, recording the identity of a figure in the image and a random generated number.

15. The search method according to claim 14, wherein the facial characteristic data comprises data representing one of a face region, the size of the face region, the characteristic of the face region, and the number of figures in the image.

16. The search method according to claim 14, wherein the characteristic region other than the face region comprises one of a dress region, a hair region, and a background region.

17. The search method according to claim 14, wherein the regional characteristic data comprises data representing the characteristic region itself and data representing regional characteristic of the characteristic region.

18. A system for searching images with figures, comprising:

an image analysis engine, comprising;

an identity analysis module, comparing figures between images of an image collection and respectively assigning an identity to each image of the image collection for each occurrence of an identified figure of the figures, the identity not derived mathematically from the pixel content of the image and relating to the identified figure; and

a representative image evaluation module, selecting and associating a representative image in the image collection for each of the identities assigned; and

an image searching interface, comprising;

a searching criterion definition module, selecting one or a plurality of the representative images as a searching criterion;

a search module, searching for images matching the searching criterion in the image collection, the matching performed in dependence upon at least the identity associated with each selected representative image; and

a display module, displaying the representative images and the images matching the searching criterion.

19. The search system according to claim 18, wherein the identity analysis module further compares figures between image groups in which figures have been respectively assigned with identities and reassigns the identities of the same figures to a unified identity.

20. The search system according to claim 18, wherein the identity analysis module comprises:

a characteristic region locating unit, locating a plurality of characteristic regions of each of the figures in the images;

a regional characteristic extraction unit, extracting a regional characteristic of the figure in each of the characteristic regions; and

a regional characteristic comparison unit, comparing the regional characteristic of each of the figures in the images and assigning an identity to the figures having the same regional characteristic.

21. The search system according to claim 20, wherein the characteristic regions comprise one of a face region, a dress region, a hair region, and a background region.

22. The search system according to claim 18, wherein the representative image evaluation module comprises:

a representative index calculation unit, respectively calculating a representative index for each of figures in images according to one of the size, the number of appearances, or the capture time of the figure; and

a representative index comparison unit, selecting the image having the largest representative index from images of the identity as the representative image of the identity in the image collection.

23. The search system according to claim 18, wherein the identity assigned by the identity analysis module comprises being recorded in a metadata corresponding to the image.

24. The search system according to claim 23, wherein the metadata comprises being stored in a header file of the image or an external file corresponding to the image.

25. The search system according to claim 18, wherein the image searching interface further comprises:

a sorting module, sorting the representative images according to the number of appearance, the region size, or the capture time of the corresponding identities, wherein the representative images are sequentially displayed by the display module.

26. The search system according to claim 25, wherein the sorting module further sorts the images of the image collection matching the searching criterion according to the number, the region size, or the capture time of figures, wherein the images matching the searching criterion are sequentially displayed by the display module.