Similar-image retrieving apparatus, similar-image retrieving...

Image analysis – Pattern recognition – Feature extraction

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

Reexamination Certificate

active

06249607

ABSTRACT:

This application is based on application No. 09-255312 filed in Japan, the contents of which are hereby incorporated by reference.
BACKGROUND OF THE INVENTION
The present invention relates to a similar-image retrieving apparatus and similar-image retrieving method for retrieving similar images in image database equipment or the like, and also relates to program storage media therefor.
In recent years, with a remarkable development in the information processing field, and particularly with improvement in performance of personal computers or the like, image database equipment and electronic filing equipment have been becoming widespread not only for business use and special use but also for general users. In these equipment, an image is captured by an input device as image data, and the captured image data is stored and managed together with additional information such as a key word, and afterwards, as required, retrieved and printed by an output device. The additional information for image data includes attribute information such as preparation date, correction date, file name and file format, and key information such as key word, shape feature, hue and speech. Then, in the image database equipment or electronic filing equipment, a desired image is retrieved by specifying the attribute information and the key information, while comparison operation of image data itself is performed for retrieving an image closely similar to a particular key image.
However, in the image database equipment or the electronic filing equipment as described above, since comparison operation of the image data itself is performed for the retrieval of a similar image, a large amount of memory resources are involved in the storage of image data, while enormous memory and processing time are required for the comparison. Thus, those equipment could not be practical systems.
This being the case, there has been proposed a similar-image retrieving apparatus in which a similar image is retrieved using edge information of images (“Retrieval of image contents, Kato and Kurita”, Journal of Information Processing Society, Vol. 33, No. 5). However, shape comparison by edge images in this similar-image retrieving apparatus has a disadvantage that the degree of extraction for external shape is so low that similar images having a rotated or translated external shape or having a scaled-up or scaled-down external shape could not be retrieved. Also, because the feature quantity of an extracted shape is represented by image data, large amounts of memories are required for storage of data, while large amounts of operations are involved in the retrieval of an image, so that long time would be required for retrieval, as a further problem.
SUMMARY OF THE INVENTION
Therefore, an object of the present invention is to provide a practical similar-image retrieving apparatus, as well as similar-image retrieving method and program storage media therefor, in which the amount of data for shape comparison can be reduced by digitizing feature quantities of external shape into numerical values while shape comparison can be achieved by simple operations using the digitized feature quantities, and moreover in which retrieval is enabled even for similar images having a rotated or translated external shape or having a scaled-up or scaled-down external shape.
In order to achieve the above-mentioned object, the present invention provides an image processing apparatus for extracting features of an image from digital image data of the image, comprising:
a detection controller for extracting an external shape of an island from an image and detecting coordinates of a pixel string positioned at an edge portion of the external shape; and
an extraction controller for extracting, as feature information of the image, coefficients determined by Fourier series expansion of the coordinates of the pixel string detected by the detection controller.
In one embodiment of the invention, the image processing apparatus further comprises
a calculation controller for comparing the feature information of the image extracted by the extraction controller with feature information of a specified image and calculating a similarity between the two images.
In one embodiment of the invention, the calculation controller calculates the similarity by multiplying coefficients for selected orders of the Fourier series by respective weights.
In one embodiment of the invention, the image processing apparatus further comprises:
a registration controller for registering in a database the feature information extracted by the extraction controller in association with the image data of the image; and
a retrieval controller for retrieving desired image data based on the feature information from among the image data registered in the database by the registration controller.
In one embodiment of the invention, the retrieval controller comprises:
an input device for receiving feature information of an image which serves as a key to retrieval; and
a calculation controller for comparing feature information of a plurality of images registered in the database with the feature information received by the input device and calculating similarities between the plurality of images and the key image so that the retrieval controller retrieves image data based on results of the calculation.
The present invention also provides an image processing method for extracting features of an image from digital image data of the image, comprising:
a detection step for extracting an external shape of an island from an image and detecting coordinates of a pixel string positioned at an edge portion of the external shape; and
an extraction step for extracting, as feature information of the image, coefficients determined by Fourier series expansion of the coordinates of the pixel string detected by the detection step.
The present invention further provides a computer program product in a memory for executing image processing to extract features of an image from digital image data of the image, comprising:
a detection control for extracting an external shape of an island from an image and detecting coordinates of a pixel string positioned at an edge portion of the external shape; and
an extraction control for extracting, as feature information of the image, coefficients determined by Fourier series expansion of the coordinates of the pixel string detected by the detection control.


REFERENCES:
patent: 5325449 (1994-06-01), Burt et al.
patent: 5640468 (1997-06-01), Hsu
patent: 6014468 (2000-01-01), McCarthy et al.
patent: 60-157681 (1985-08-01), None
patent: 5-28266 (1993-02-01), None
patent: 5-274372 (1993-10-01), None
patent: 6-274193 (1994-09-01), None
patent: 6-295318 (1994-10-01), None
patent: 7-271949 (1995-10-01), None
patent: 8-16789 (1996-01-01), None
patent: 8-329096 (1996-12-01), None
The Transaction of the Japanese Information Processing Society, vol. 33, No. 5.; May 1992.
Fundamentals of Image Recognition (II); Apr. 20, 1990.
Dr. Gupta, “Visual Information Retrieval Technology A Virage Perspective”, Virage Image Engine API Specification, Feb. 1997.

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Similar-image retrieving apparatus, similar-image retrieving... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Similar-image retrieving apparatus, similar-image retrieving..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Similar-image retrieving apparatus, similar-image retrieving... will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-2441568

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.