METHOD FOR IMAGE ANALYSIS, ESPECIALLY FOR MOBILE STATIONS

• US 20100296729A1
• Filed: 10/28/2008
• Published: 11/25/2010
• Est. Priority Date: 11/05/2007
• Status: Active Grant

• Current Assignee: T-Mobile International AG (Deutsche Telekom AG)
• Original Assignee: T-Mobile International AG (Deutsche Telekom AG)
• Inventors: Mossakowski, Gerd
• Granted Patent: US 8,532,389 B2
• Time in Patent Office: Days
• US Class Current: 382/165
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/18076 by analysing connectivity, e.g. edge linking, connected component analysis or slices

Abstract

A robust OCR system requiring little computing capacity is obtained by first carrying out an adaptive pre-processing optimised in terms of pixel groups, which analyses the image in line segments. The most significant difference compared to previously known methods is that there is no longer a direct pattern comparison, instead the line segments are gone over in as optimum a manner as possible. The corresponding character is then deduced from the sequence of movements. As this sequence of movements can be scaled well and described in a relatively simple manner, this technique is especially suitable for mobile use. The sequence of movements of know characters is stored in a search word, such that the letters can be directly deduced from the movement. A dictionary/lexicon can also be used. If words are recognised by means of the dictionary/lexicon, the recognised letters can be used for an even more optimised character font identification. The invention is advantageous in that a robust OCR system is provided, which also requires little computing capacity. The system according to the invention is robust especially in that the recognition works better than with conventional systems even under bad conditions, especially light ratios and interferences.

Claims

1. A method for OCR acquisition which carries out the following steps:

a) recognition of line segments by pixel group oriented list formation, wherein the lists each represent individual line segments;

b) tracing of the letters on the basis of the generated lists;

c) comparison of the sequence of movements in the case of the retracing of the letters with standardized reference letters, stored in a solution tree.

5. The method according to claim 1, characterized in that lists of similar pixel groups are formed.

17. The method according to claim 1, characterized in that a dictionary/lexicon is used, with the help of which the recognized letters are used for an even more optimized character font recognition.

18. The method according to claim 1, characterized in that the recognized words are translated into a selectable language and optically and/or acoustically output.

19. The method according to claim 1, characterized in that through acknowledgments of recognized words solution trees and line segment widths of the prototype can be correspondingly optimized.

20. The method according to claim 1, characterized in that the running determination and output of the pixel groups sorted by priorities takes place by means of a used image recording system, in particular a canner or CCD camera integrated in a mobile telephone.

2. A method for the analysis of image data which consist of an array of individual pixels, wherein each pixel exhibits a chronologically changing pixel value which describes the color or brightness information of the pixel, wherein the following steps are carried out:

a) the determination of a priority value for each pixel of the array by fixing the used pixel as a reference pixel and calculation of a pixel difference value with the help of the current pixel value of the reference pixel with relation to the current pixel values of a previously defined group of adjacent pixels;

b) combination of the pixels used for the calculation of the priority value into one pixel group,

c) sorting the pixel groups with the help of the priority value of the allocated reference pixel and storage in a priority array;

d saving and/or transferal of the pixel groups according to their priority in the priority array, wherein for optimization of the computing capacity only a portion of the pixel groups are used for formation of the lists, characterized in that in addition a position factor is flowed into the priority value which is all the greater, the closer the pixel group is to a predefined start pixel dependent on the language.

3. The method according to claim 2, characterized in that the pixel difference value results from the difference of the pixel value of a considered pixel to the pixel value of some of its considered neighboring pixels of the pixel group.

4. The method according to claim 2, characterized in that the pixel difference value permits inferences about the line segment width.

6. The method according to claim 2, characterized in that after Steps 1a) through 1d) the following steps are carried out:

an adaptive pre-processing optimized in terms of pixel groups is carried out, which analyzes the image in line segments, wherein subsequently the line segments are gone over in as optimum a manner as possible, wherein the corresponding character is then deduced from the sequence of movements via stored search words/solution trees.

7. The method according to claim 2, characterized in that after Steps 1a) through 1d) the following steps are carried out:

similar pixel groups are each compiled in a separate list and each list thus gained is in the process sorted in such a way that the pixel groups which exhibit a lower Y position are sorted in descending order, wherein if several similar pixel groups lie at identical Y positions, new lists are generated for said pixel groups, wherein vectors are derived from these lists and the pixel groups with the lowest and the highest Y value are selected and wherein between these pixel group positions a line is calculated and wherein the deviations of the other pixel groups to this line are determined.

8. The method according to claim 7, characterized in that in case all the deviations lie below a specified threshold value, a description vector is found for this list, but if the deviations lie above a threshold value, the list is divided and an attempt is made to generate corresponding vectors for each sublist.

9. The method according to claim 8, characterized in that the list is divided where the greatest deviations to the calculated line are present.

10. The method according to claim 7, characterized in that the vector list vector list is subsequently standardized e.g. to maximum Y difference.

11. The method according to claim 10, characterized in that the standardized vector list passes through a solution tree in which the different letters are stored.

12. The method according to claim 7, characterized in that vectors touching one another are combined in a further vector list and the Y values are correspondingly sorted.

13. The method according to claim 7, characterized in that the width of the used pixel group is selected in such a way that it is three times the line width and the optimum height of the pixel group is dependent on the font height.

14. The method according to claim 7, characterized in that the image is then scanned again with the pixel groups thus optimized.

15. The method according to claim 7, characterized in that optimized result trees are generated for each text with this font.

16. The method according to claim 7, characterized in that for typewritten text letters or even syllables already recognized are stored as pixel group masters.