Active image tagging

• US 8,825,744 B2
• Filed: 06/10/2010
• Issued: 09/02/2014
• Est. Priority Date: 06/10/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Wang, Meng, Hua, Xian-Sheng, Yang, Kuiyuan
• Primary Examiner(s): Fan, Hua
• Application Number:
  US12/813,328
• Publication Number:
  US 20110307542A1
• Time in Patent Office: 1,545 Days
• Field of Search: 709/203
• US Class Current: 709/203
• CPC Class Codes: G06F 16/5866 using information manually generated, e.g. tags, keywords, comments, manually generated location and time information

Abstract

Methods and systems for active image tagging are usable to build large datasets of tagged images by combining manual tagging by a user and automatic tagging by a computing device based on the manual tagging. Such tags may be used to effectively sort, organize, link, and search for images within large datasets of images. Additionally, the active image tagging may be configured to utilize a tagging game where multiple users manually tag images by playing a game on a computing device.

Claims

1. A computer-implemented method for active tagging, comprising:

performed by one or more processors executing computer-readable instructions;

selecting a file from a data set stored in a memory based at least in part on an informativeness measure of the file, wherein the informativeness measure is based at least in part on at least one of a citation measure, an ambiguity measure, or a diversity measure;

providing the file to a client device for manual tagging by a user of the client device;

receiving a manually tagged file from the user of the client device;

storing the manually tagged file in the data set;

predicting a tag for another file based at least in part on the manually tagged file to create another tagged file in the data set that is distinct from the manually tagged file in the data set; and

storing the other tagged file in the data set.

2. The computer-implemented method of claim 1, wherein the manually tagged file and the other tagged file comprise image files.

3. The computer-implemented method of claim 1, wherein selecting a file from the data set comprises selecting the file from a subset of data in the data set.

4. The computer-implemented method of claim 1, further comprising iteratively performing the method until no untagged files exist in the data set.

5. The computer-implemented method of claim 3, further comprising iteratively performing the selecting the file, the providing the file, the receiving the manually tagged file from the user, and the storing the manually tagged file until no untagged files exist in the subset of data.

6. The computer-implemented method of claim 5, further comprising iteratively performing the predicting the tag for another file and the storing the other tagged file until no untagged files exist in the data set, subsequent to iteratively performing the selecting the file, the providing the file, the receiving the manually tagged file, and the storing the manually tagged file until no untagged files exist in the subset of data.

7. The computer-implemented method of claim 3, wherein the selecting a file from the data set comprises selecting the file from the subset of data based at least in part on the informativeness measure of the file.

8. The computer-implemented method of claim 1, wherein the citation measure defines a number of files neighboring the file, the ambiguity measure defines a similarity between a tag of each neighboring file and a tag of the file, and the diversity measure defines a difference between a tag of each file in a sample and the tag of the file.

9. The computer-implemented method of claim 1, wherein providing the file to a client device for manual tagging by the user comprises providing the file and a text entry field associated with the file to the client device for displaying to the user.

10. The computer-implemented method of claim 1, wherein providing the file to a client device for manual tagging by the user comprises transmitting the file through a network interface to the client device connected to a network.

11. One or more computer-readable storage media, storing processor-executable instructions that, when executed on a processor, perform acts for active tagging, the acts comprising:

hosting a file tagging game;

selecting a file stored in a memory, wherein the selecting comprises analyzing one or more tags to determine an appearance probability related to the selected file;

providing the selected file to at least a first client device and a second client device of the file tagging game;

receiving a game-defined tag associated with the selected file from at least the first client device and the second client device executing the file tagging game;

in response to determining that the game-defined tag received from the first client device matches a game-defined tag received from the second client device, storing the game-defined tag with the selected file in the memory;

predicting a tag for another file in the memory based at least in part on the selected file and the game-defined tag stored in the memory;

associating the tag with the other file to create another tagged file; and

storing the other tagged file in the memory.

12. The one or more computer-readable storage media of claim 11, wherein the file and the other file comprise image files.

13. The one or more computer-readable storage media of claim 11, wherein the performed acts further comprise iteratively performing the method until no untagged files exist in the memory.

17. The computer-implemented method of claim 11, wherein providing the selected file to the at least first client device and the second client device further comprises providing the selected file and a text entry field associated with the selected file to the at least first client device and the second client device for display.

18. The computer-implemented method of claim 11, wherein providing the selected file to the at least first client device and the second client device further comprises transmitting the file through a network interface to the at least first client device and second client device connected to a network.

19. The computer-implemented method of claim 11, wherein the selecting a file stored in a memory, further comprises selecting the file from a subset of data in a data set.

20. The computer-implemented method of claim 19, further comprising iteratively performing the selecting the file stored in the memory until no untagged files exist in the subset of data in the data set.

14. A system for iterative active image tagging comprising:

memory and a processor;

an image file selection module, stored in the memory and executable on the processor, that selects an image file from a subset of data stored in the memory based at least in part on an informativeness measure comprising at least one of a citation measure defining a number of neighboring image files to the image file, an ambiguity measure defining a similarity between a tag of the image file and tags of one or more of the neighboring image files, and a diversity measure defining a difference between a tag of each image file in a sample and the tag of the image file;

a manual image file tagging module, stored in the memory and executable on the processor, that provides the selected image file to a plurality of client devices for a plurality of users to manually tag, wherein a manually tagged image file is stored in the memory in response to determining that the plurality of users select a same tag for the selected image file;

an automatic image file tagging module, stored in the memory and executable on the processor, that provides automatic image tags to other image files in the subset of data stored in the memory based at least in part on the manually tagged image file; and

an iteration module, stored in the memory and executable on the processor, that iteratively activates the manual image file tagging module and the automatic image file tagging module until tagging is complete.

15. The system of claim 14, wherein the manual image file tagging module also provides the selected image file to another client device for a user of the other client device to create the manually tagged image file to be stored in the memory, the client device and the other client device being connected through a tagging game and the creation of the manually tagged image file being based at least in part on a result of the tagging game.

16. The system of claim 14, wherein the image file selection module also selects the image file based at least in part on an informativeness measure of the file.