Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-08-02
2002-10-08
Homere, Jean R. (Department: 2177)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C382S299000
Reexamination Certificate
active
06463432
ABSTRACT:
This application is based on application Nos. 10-218854, 10-231757 and 10-255232 filed in Japan, the contents of which are hereby incorporated by reference.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a technique of comparing an image with another image, and more particularly to a technique of retrieving an image similar to a specified image (hereinafter referred to as “key image”) from a plurality of images stored in a database to be searched. Further, the present invention relates to a technique of sequentially displaying a plurality of images while searching.
2. Description of the Background Art
With recent remarkable development in information processing and in particular, improvement in performance of personal computer, image database apparatus, electronic filing apparatus and the like become widespread for general user (consumer) as well as specific application such as business area. The image database apparatus and the electronic filing apparatus capture images and documents through an input device and store the captured data associated with added information such as keyword into a database as image data or document data, manage the database, search the database for images or documents as required and display or print the retrieved data on an output device.
Without appropriate added information such as keyword associated with the image data in the image database storing a large amount of data, however, an effective search is impossible and hence difficulty in quick retrieval of a desired image results in poor practicality.
In such a search for images, one of techniques for searching a database storing a plurality of images for an image similar to a key image adopts a method of comparing feature values representing features of images. There are various types of feature values. For example, as the feature values of color and texture, chromaticness (saturation) and the edge intensity are obtained, respectively, and an image having values close to those of the key image is retrieved from a plurality of images. As to this technique, for example, a retrieval software of Exculibur named as “Visual RetrievalWare” and a retrieval software of Virage named as “VIR Image Engine” are well known.
In these well-known similar-image retrieval technique, a fixed default value or a user-set value is used as a weighting factor for each feature value and a comparison is made on the basis of the weighting factor for each feature value. If a search is made with the weighting factor for a feature value increased or decreased and the search result is unsatisfactory, a further search is required with the weighting factor for another feature value increased or decreased.
For this reason, in order to perform a search for a desired image, it is desirable for a user to know in advance which feature value of the key image should be emphasized in the search (in other words, which feature value is given a large weighting factor).
Further, these well-known similar-image retrieval technique do not consider resolution of the images to be compared with each other. It is obvious that feature values of an image to be extracted change with variation of its resolution. For example, in an image with high resolution, fine edges are clearer and the feature value of texture is larger. Various colors are found independently in an image with high resolution while mixed in that with low resolution, and the images with high resolution and low resolution have different feature values of each color.
When an image search is performed while the key image and an image to be compared are displayed on a screen for observation, these images are reduced by limitation of the size of the screen to have lower resolution in some cases. Even in this situation, a similar-image retrieval with resolution of original image is required in some cases and a similar-image retrieval with impression of the image on the browser is required in other cases.
On the other hand, regardless of whether appropriate added information is associated with the image data or not, studied is an effective use of image data in a case where a large amount of data are stored in the image database.
One of effective uses of image data is a method of sequentially displaying images at a constant time interval (“slide show”), disclosed, for example, in Japanese Patent Examined Publication No. 4-17545, Japanese Patent Examined Publication No. 4-24747 and so on. The above publications disclose a function for a variable setting of the display interval of the slide show.
Since the images, however, are displayed only in a predetermined order such as the order of addresses at which the image data are stored in the background-art method of sequentially displaying images, it is impossible to make a sequential display in consideration of the contents of images and the search is made with very low efficiency.
SUMMARY OF THE INVENTION
According to an aspect of the present invention, an apparatus for comparing a plurality of objective images with a key image comprises: feature-value calculation means for obtaining a plurality of feature values for each of the objective images and the key image; weighting-factor setting means for setting weighting factors for the plurality of feature values respectively from the plurality of feature values of the key image; similarity-value calculation means for obtaining similarity values of each of the plurality of objective images to the key image with respect to the feature values; and total similarity-value calculation means for obtaining a total similarity value of each of the plurality of objective images to the key image from the similarity values and the weighting factors.
In the apparatus of this aspect, the weighting factors with respect to feature values specific to the key image can be obtained on the basis of the feature values of the key image while the similarity values of each objective image to the key image can be obtained with respect to feature values. Therefore, since the total similarity value can be obtained on the basis of the weighting factors specific to the feature values of the key image and the similarity values of the objective image to the key image with respect to the feature values, it is possible to compare the objective image with the key image without user's setting of the weighting factors by trial and error.
According to another aspect of the present invention, a similar-image retrieval apparatus comprises: resolution converting means for obtaining converted images with different resolutions for each of a plurality of objective images and a key image; feature-value calculation means for obtaining feature values with respect to each of the different resolutions; and similar-image retrieval means for obtaining a similar image out of the plurality of objective images on the basis of feature values for the key image with a first resolution and feature values of the plurality of objective images with a second resolution common to the first resolution.
In the similar-image retrieval apparatus, when judgment on whether two images are similar or not is performed on the basis of the feature values of the key image and the objective image, the feature values in accordance with a resolution of the images to be compared are adopted. Therefore, even if the key image and the objective image are reduced by limitation of the size of a screen to have lower resolution when these images are too large to be displayed, it is possible to perform a judgment on similarity with the resolution of the images before being reduced. On the other hand, by performing the judgment with the resolution of displayed images, a retrieval result and an impression of observed images are matched.
According to still another aspect of the present invention, an image sequential display apparatus comprises: a memory for storing a plurality of image data; image feature-value extraction means receiving the plurality of image data from the memory, for extracting image feature values for each of the plurali
Homere Jean R.
McDermott & Will & Emery
Minolta Co. , Ltd.
Wassum Luke S
LandOfFree
Apparatus for and method of retrieving images does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus for and method of retrieving images, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus for and method of retrieving images will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2921285