Anonymizing textual content in image training data
A computer-implemented method for modifying image data, the method including: loading unmodified image data; detecting at least two alphanumeric characters in an image represented by the loaded image data; selecting one or more of the detected alphanumeric characters, wherein the number of selected alphanumeric characters is smaller than the total number of detected alphanumeric characters; modifying the loaded image data by removing one or more character sections of the loaded image data, wherein each character section corresponds to an area of a selected alphanumeric character; storing the modified image data. Also provided is a corresponding data processing system, computer program product and computer-readable storage medium.
1. A computer-implemented method for modifying image data, the method comprising:
loading unmodified image data;
detecting at least two alphanumeric characters in an image represented by the loaded image data;
selecting one or more of the detected alphanumeric characters, wherein the number of selected alphanumeric characters is smaller than the total number of detected alphanumeric characters;
modifying the loaded image data by removing one or more character sections of the loaded image data, wherein each character section corresponds to an area of a selected alphanumeric character;
modifying the loaded image data by shuffling one or more character sections of the loaded image data, wherein each character section corresponds to an area of an unselected alphanumeric character; and
storing the modified image data.
2. The method of claim 1 , wherein the number of selected alphanumeric characters is approximately half the number of all detected alphanumeric characters.
3. The method of claim 1 , wherein:
detecting at least one word comprising two or more alphanumeric characters in the image represented by the loaded image data,
selecting one or more of the detected alphanumeric characters of each detected word.
4. The method of claim 3 , further comprising selecting approximately half the number of detected alphanumeric characters of each word.
5. The method of claim 1 , further comprising:
modifying the loaded image data by replacing at least one of the removed character sections of the loaded image data with a character section of the loaded image data corresponding to an area of an unselected alphanumeric character.
6. The method of claim 3 , further comprising:
modifying the loaded image data by replacing at least one of the removed character sections of the loaded image data belonging to the at least one word with a character section of the loaded image data corresponding to an area of an unselected alphanumeric character belonging to the same word as the removed character section.
7. The method of claim 3 , wherein shuffling of character sections corresponding to alphanumeric characters belonging to the at least one word is limited to shuffling within the same word.
8. The method of claim 1 , further comprising:
detecting at least one face in an image represented by the loaded image data;
modifying the loaded image data by removing one or more portrait sections of the loaded image data, wherein each portrait section corresponds to an area of a detected face.
9. The method of claim 8 , wherein removing one or more portrait sections of the loaded image data comprises replacing at least one removed portrait section with a blurred version of the same portrait section.
10. The method of claim 8 , further comprising:
detecting at least one additional face in an image represented by the loaded image data using the at least one detected face as a template;
modifying the loaded image data by removing one or more additional portrait sections of the loaded image data, wherein each additional portrait section corresponds to an area of a detected additional face.
11. The method of claim 10 , wherein removing one or more additional portrait sections of the loaded image data comprises replacing at least one additional portrait section with a blurred version of the same additional portrait section.
12. The method of claim 1 , further comprising:
detecting at least one written signature in an image represented by the loaded image data;
modifying the loaded image data by removing one or more signature sections of the loaded image data, wherein each signature section corresponds to an area of a detected written signature.
13. The method of claim 12 , wherein removing one or more signature sections of the loaded image data comprises replacing at least one removed signature section with a blurred version of the same signature section.
14. The method of claim 1 , further comprising:
detecting at least one machine-readable code in an image represented by the loaded image data;
modifying the loaded image data by removing one or more code sections of the loaded image data, wherein each code section corresponds to an area of a detected machine-readable code.
15. The method of claim 14 , wherein removing one or more code sections of the loaded image data comprises replacing at least one removed code section with a blurred version of the same code section.
16. A data processing system comprising means for carrying out the method of claim 1 .
17. A computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of claim 1 .
18. A computer-implemented method for modifying image data, the method comprising:
loading unmodified image data;
detecting at least two alphanumeric characters in an image represented by the loaded image data;
selecting one or more of the detected alphanumeric characters, wherein the number of selected alphanumeric characters is smaller than the total number of detected alphanumeric characters;
detecting at least one word comprising two or more alphanumeric characters in the image represented by the loaded image data;
selecting one or more of the detected alphanumeric characters of each detected word;
selecting approximately half the number of detected alphanumeric characters of each word;
modifying the loaded image data by removing one or more character sections of the loaded image data, wherein each character section corresponds to an area of a selected alphanumeric character; and
storing the modified image data.
19. The method of claim 18 , further comprising:
modifying the loaded image data by replacing at least one of the removed character sections of the loaded image data belonging to the at least one word with a character section of the loaded image data corresponding to an area of an unselected alphanumeric character belonging to the same word as the removed character section.
20. The method of claim 18 , wherein the number of selected alphanumeric characters is approximately half the number of all detected alphanumeric characters.