DATA PROCESSING APPARATUS, COMPUTER PROGRAM PRODUCT, AND DATA PROCESSING METHOD

• US 20090074288A1
• Filed: 09/16/2008
• Published: 03/19/2009
• Est. Priority Date: 09/19/2007
• Status: Active Grant

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: NISHIDA, Hirobumi
• Granted Patent: US 8,254,669 B2
• Time in Patent Office: Days
• US Class Current: 382/159
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
• CPC Class Codes: G06V 10/70 using pattern recognition or machine learning optical pattern recognition or electronic computations therefor G06V10/88
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/18143 Extracting features based on salient regional features, e.g. scale invariant feature transform [SIFT] keypoints
• CPC Class Codes: G06V 30/413 Classification of content, e.g. text, photographs or tables

Abstract

A data processing apparatus includes a feature-value calculating unit that calculates an image feature value indicating a feature of image data, a case database including a case set including a correspondence of image feature values and functions, and an optimum-function predicting unit that predicts an optimum function based on the case database and the image feature value calculated by the feature-value calculating unit. Due to the optimum-function predicting unit, work efficiency of a user can be improved.

Claims

1. A data processing apparatus comprising:

a feature-value calculating unit that calculates an image feature value indicating a feature of image data;

a case database including a case set including a correspondence of image feature values and functions; and

an optimum-function predicting unit that predicts an optimum function based on the case database and the image feature value calculated by the feature-value calculating unit.

2. The data processing apparatus according to claim 1, further comprising a predictor building unit that builds a predictor that predicts an optimum function based on the case database, wherein

the optimum-function predicting unit predicts an optimum function by using the predictor.

3. The data processing apparatus according to claim 2, wherein the optimum-function predicting unit includes

a distance calculating unit that calculates a distance of a feature value vector between a prototype point and a target point of the image data, the prototype point being from the case set and to be used in the predictor;

and

a unit that outputs an optimum classification category that is a classification category corresponding to class information appended to a prototype point at which the distance is minimized.

4. The data processing apparatus according to claim 2, further comprising:

an input receiving unit that receives an input of a function from a user; and

an updating unit that updates, when the optimum function is different from the function received from the user, the case set to a sum set of a new case set from the user and the case set, wherein

the predictor building unit includes a sequential building unit that rebuilds, every time the case set is updated by the updating unit, a predictor using updated case set.

5. The data processing apparatus according to claim 3, wherein the predictor building unit includes a batch building unit that rebuilds, upon judging that a predetermined series of works is finished, the predictor using all case sets in the case database.

6. The data processing apparatus according to claim 4, wherein the optimum-function predicting unit includes

a distance calculating unit that calculates a distance of a feature value vector between prototype data and the image data, the prototype data being from among the case set and to be used in the predictor;

and

a unit that outputs an optimum function that is a function corresponding to class information appended to the prototype data with which the distance is minimized.

7. The data processing apparatus according to claim 4, wherein the updating unit includes

a unit that detects a set of pieces of case data having feature value vectors of substantially same values and indicating different functions among case sets in the case database;

and

a unit that excludes temporally old one of the case data from the case set and incorporates temporally new one of the case data in the case set.

8. The data processing apparatus according to claim 4, wherein the sequential building unit includes a parameter updating unit that updates, every time new image data is input from a user, a parameter of a distance function for prototype data having a feature value close to an image feature value calculated from the new image data among case data in the case set.

9. The data processing apparatus according to claim 8, wherein the parameter updating unit updates weight, the weight being determined based on the case data, such that a prediction error is minimized, the prediction error being calculated from only prototype data having an image feature value close to the new image data.

10. The data processing apparatus according to claim 5, further comprising an input receiving unit that receives an input of a function from a user, wherein

the batch building unit includes a supplementing unit that reduces, when the optimum function is different from the function received from the user, a volume of the case set to be as small as possible and optimizes weight of the prototype data such that a prediction error is minimized.

11. The data processing apparatus according to claim 10, wherein the supplementing unit includes

a unit that adds the new case data to the case set and deletes case data contradictory to the new case data from the case set;

and

a unit that updates weight of the prototype data for a near data group of the new case data.

12. The data processing apparatus according to claim 1, further comprising:

a weight calculating unit that calculates a weight for each of cases in the case set; and

a case-set building unit that builds, in the case data base, a new case set in which number of cases is reduced based on the weight, wherein

the optimum-function predicting unit predicts an optimum function based on the new case set, the weights, and the image feature value.

13. The data processing apparatus according to claim 12, wherein the optimum-function predicting unit includes

a distance calculating unit that calculates, for each of cases in the new case set, a weighted distance of a feature value vector to the image data;

and

a unit that outputs, as an optimum function from the new case set, a case in which the distance is minimized.

14. The data processing apparatus according to claim 1, further comprising:

an area extracting unit that executes area extraction on the image data by using at least two area extraction systems;

an evaluating unit that ranks each of extracted areas obtained by the area extraction systems and evaluates the area extraction systems;

a feature-value-data recording unit that records, as supervised feature value data, the image feature value and a best one of the area extraction systems based on evaluation by the evaluating unit; and

a learning unit that learns a selection rule for selecting the best area extraction system from the image feature value using the supervised feature value data, wherein

the data processing apparatus builds the case database.

15. The data processing apparatus according to claim 14, wherein the feature-value calculating unit includes

a block dividing unit that exclusively divides the image data into rectangular blocks;

a block classifying unit that classifies the blocks into one of predetermined elements that form the image data; and

a calculating unit that calculates an image feature value of the image data based on a result of classification of the blocks.

16. The data processing apparatus according to claim 15, wherein the block classifying unit includes

an image generating unit that generates a plurality of images having different resolutions from the blocks;

a feature-value-vector calculating unit that calculates a feature value vector from each of the images; and

a classifying unit that classifies the blocks into one of predetermined elements based on the feature value vector.

17. The data processing apparatus according to claim 16, wherein the feature-value-vector calculating unit includes

a binarizing unit that binarizes each of the images;

a pixel-feature calculating unit that calculates a feature for each of pixels of each of binarized images based on values of each of the pixels and a pixel corresponding to a local pattern formed of neighboring pixels near each of the pixels; and

an adding unit that adds up, over the entire image, features calculated for the pixels.

18. The data processing apparatus according to claim 16, wherein the feature-value-vector calculating unit includes

a pixel-feature calculating unit that calculates a feature for each of pixels of the images based on values of each of the pixels and a pixel corresponding to a local pattern formed of neighboring pixels near each of the pixels;

and

an adding unit that adds up, over the entire image, features calculated for the pixels.

19. A computer program product comprising a computer usable medium having computer readable program codes embodied in the medium that, when executed, causes a computer to execute:

calculating an image feature value indicating a feature of image data; and

predicting an optimum function based on a case database, the case data base including a case set including a correspondence of image feature values and functions, and the image feature value calculated at the calculating.

20. A data processing method executed by a data processing apparatus including a control unit and a storing unit, the data processing method comprising:

calculating an image feature value indicating a feature of image data; and

predicting an optimum function based on a case database, the case data base including a case set including a correspondence of image feature values and functions, and the image feature value calculated at the calculating.