JPH064600A - Image search method and image search apparatus - Google Patents

Abstract

(57) [Summary] [Purpose] To provide an image search method and an image search device capable of easily creating, inputting, and correcting image information for search. [Structure] The line figure feature extraction unit 2 extracts line figure features that characterize the image from image information. The image storage unit 1 stores the image information and the line figure feature extracted by the line figure feature extraction unit 2 in association with each other. The online input feature extraction unit 3 extracts online input features from the online input data. The online input feature storage unit 4 stores the online input features extracted by the online input feature extraction unit 3. The feature retrieval unit 5
The similarity between the line figure feature stored in the image storage unit 1 and the online input feature stored in the online input feature storage unit 4 is judged to search the image information corresponding to the online input data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image retrieving method and an image retrieving apparatus for retrieving pre-collected image information by using image features themselves without using specially added code information. Is.

[0002]

2. Description of the Related Art In recent years, with the progress of memory technology and database technology, it has become widespread to accumulate and retrieve a large amount of image information. For this search, a method is generally used in which code information such as a number, a title, or a keyword is added to image information at the time of input, and desired image information is obtained by searching the code information. However, in addition to that, the method of searching from the characteristics themselves such as the shape of image information does not require special addition of code information at the time of input, and it is easy to search for multiple image information that are similar in image itself. Is becoming influential.

In addition, with the development of pen input tablets and liquid crystal display technology, it has become easy to input graphic images online. As shown in FIG. 3, the conventional image retrieval apparatus as described above includes an image storage unit 31.
And the line figure feature extraction unit 32 and the feature search unit 33. The image storage unit 31 stores the input / accumulated image information and the line graphic feature associated with the image information. The line figure feature extraction unit 32 extracts line figure features that characterize the image information from the input image information. The feature search unit 33 determines the similarity between the line figure features and outputs the search result.

In this conventional image retrieval apparatus, the accumulated image information is first stored in the image storage section 31 and passed to the line figure feature extraction section 32. Accordingly, the line figure feature extraction unit 32 performs a process such as edge detection process to extract the line figure feature that characterizes the image, and stores it in the image storage unit 31 in association with the image information. When performing a search, the searcher creates a search image in advance and inputs it to the line figure feature extraction unit 32. As a result, the line figure feature extraction unit 32 performs processing such as edge detection processing on the search image to extract line figure features that characterize the image, and the feature search unit 33.
Pass to. As a result, the feature search unit 33 determines the similarity with the line graphic feature stored in the image storage unit 31 and associated with the accumulated image information, and determines the similarity with the line graphic feature of the search image. Image information associated with a plurality of line figure features is output as a search result in order from the highest or the highest.

[0005]

However, in the above-mentioned conventional configuration, since the image information itself similar to the image information to be searched for must be created by some means and input from a scanner or other device to be processed, the search is performed. There is a problem that it takes time to create and input the image information to be used. In addition, there is a problem that it is difficult to understand to what extent similar search image information needs to be created, and it is troublesome to correct and re-enter the search image information.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an image search method and an image search apparatus capable of easily creating, inputting or correcting image information for search.

[0007]

According to the invention of claim 1, a line graphic feature extracting step for extracting a line graphic feature characterizing the image from image information, image information, and the line graphic feature extracting step. An image storing step of storing line figure features in association with each other, an online input feature extracting step of extracting online input features from online input data, and an online input feature storing online input features extracted in the online input feature extracting step. A storing step; a feature searching step of searching the image information corresponding to the online input data by determining the similarity between the line figure feature stored in the image storing step and the online input feature stored in the online input feature storing step. Run It is characterized in that.

According to a second aspect of the present invention, a line figure feature extraction unit for extracting a line figure feature that characterizes the image from the image information, the image information, and the line figure feature extracted by the line figure feature extraction unit are associated with each other. An image storage unit that stores the stored image, an online input feature extraction unit that extracts online input features from online input data, an online input feature storage unit that stores the online input features extracted by the online input feature extraction unit, and the image And a feature search unit for searching image information corresponding to online input data by determining similarity between the line figure feature stored in the storage unit and the online input feature stored in the online input feature storage unit. It is characterized by that.

[0009]

According to the first aspect of the present invention, the line figure feature extracting step extracts the line figure features that characterize the image from the image information, and the image storing step extracts the image information and the line figure feature extracting step. The line figure features are stored in association with each other, the online input feature extraction step extracts online input features from the online input data, and the online input feature storage step stores the online input features extracted in the online input feature extraction step. , In the feature search step,
Image information corresponding to the online input data is searched by determining the similarity between the line figure feature stored in the image storing step and the online input feature stored in the online input feature storing step.

According to the second aspect of the invention, the line figure feature extraction unit extracts line figure features that characterize the image from the image information. The image storage unit stores the image information in association with the line figure feature extracted by the line figure feature extraction unit. The online input feature extraction unit extracts online input features from the online input data. The online input feature storage unit stores the online input features extracted by the online input feature extraction unit. The feature search unit determines the similarity between the line figure feature stored in the image storage unit and the online input feature stored in the online input feature storage unit and retrieves image information corresponding to the online input data.

[0011]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a block diagram of an image search apparatus according to an embodiment of the present invention. The image search apparatus includes an image storage unit 1, a line figure feature extraction unit 2, an online input feature extraction unit 3, and an online input feature storage. It includes a unit 4 and a feature search unit 5. The image storage unit 1 stores the input / accumulated image information and the line graphic feature associated with the image information. The line figure feature extraction unit 2
A line figure feature that characterizes the image information is extracted from the image information. The online input feature extraction unit 3 extracts online input features that characterize the online input data from the search online input data. The online input feature storage unit 4 stores online input features. The feature search unit 5 determines the similarity between the line figure feature and the online input feature and outputs the search result.

Next, the operation will be described. The accumulated image information is first stored in the image storage unit 1 and passed to the line figure feature extraction unit 2. As a result, the line figure feature extraction unit 2
Extracts the line figure features that characterize the image and stores them in the image storage unit 1 in association with the image information. The line graphic feature is data that can be easily compared with the online input feature later, for example, data such as vector information in which the edge is vectorized after the edge is detected as shown in FIG.

When performing a search, the online input feature extraction unit 3 extracts online input features from the input online input data, and the online input feature storage unit 4
To store. As the online input feature, generally, a series of coordinate point sequence from pen down to pen up (hereinafter referred to as “stroke”) when input by a pen can be used. When a certain amount of online input features are stored in the online input feature storage unit 4,
According to the searcher's instruction, the feature search unit 5 causes the online input feature stored in the online input feature storage unit 4,
The similarity to the line figure feature stored in the image storage unit 1 in association with the image information is determined, and the line image feature of the search image has the highest similarity or a plurality of lines in order from the highest one. The image information associated with the graphic feature is output as a search result. At this time,
It is not necessary to use the direction and order of the online input features to judge the similarity, but in order to cope with continuous drawing and disconnection of online input, the number of adjacent line drawing features and online input features It is necessary to perform a process of judging similarity by trying to integrate these elements.

The online input feature extraction unit 3 has a function of deleting or modifying individual online input features (for example, one of strokes) from the online input feature storage unit 4 in order to enable correction of the online input features. have. Further, the feature search unit 5 has a function of outputting the search result once according to the searcher's instruction, and then adding another online input feature for narrowing down the result or the like, and determining the similarity again. This allows you to first do a rough search and then make additional online inputs to continue a more detailed search.

As described above, the line graphic feature extraction unit 2 for extracting the line graphic feature that characterizes the image from the image information, the image information, and the line graphic feature extracted by the line graphic feature extraction unit 2 are stored in association with each other. An image storage unit 1, an online input feature extraction unit 3 that extracts online input features from online input data, an online input feature storage unit 4 that stores the online input features extracted by the online input feature extraction unit 3, and an image A feature search unit 5 that searches the image information corresponding to the online input data by determining the similarity between the line figure feature stored in the storage unit 1 and the online input feature stored in the online input feature storage unit 4. The feature search unit 5 is equipped with
After outputting any number of image information corresponding to the online input data as a search result, the search is continued by combining the online input data that is continuously input and the online input data that has been input so far. Sometimes it is possible to search for any number of pieces of image information that are similar to each other without the need to add code information, but the characteristics of the image for search can be easily input as online input characteristics. The image information similar to the online input features already input at the time of can be easily searched, and the features of the image for search can be easily corrected by adding or partially modifying the stored online input features. -You can re-enter and easily search again.

[0016]

As described above, according to the present invention, the line figure feature that characterizes the image is extracted from the image information, and the image information and the line figure feature are stored in association with each other. Is extracted, the online input features are stored, the similarity between the line figure features and the online input features is determined, and the image information corresponding to the online input data is searched.Therefore, without adding code information at the time of input, In an image search device capable of searching image information whose images are similar to each other, it is possible to easily create, input or correct image information for search.

Further, after the feature retrieval unit outputs any number of image information corresponding to the online input data as a retrieval result, the online input data continuously input and the online input data previously input are combined. If the configuration is such that the search is continued, it is possible to perform a rough search first and then perform additional online input to continue the more detailed search.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an image search device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a state of extracting line figure features.

FIG. 3 is a configuration diagram of a conventional image search device.

[Explanation of symbols]

 1 image storage unit 2 line figure feature extraction unit 3 online input feature extraction unit 4 online input feature storage unit 5 feature search unit

Claims (3)

[Claims] 1. A line figure feature extraction step for extracting a line figure feature that characterizes the image from image information, and an image storage for storing the image information and the line figure feature extracted in the line figure feature extraction step in association with each other. An online input feature extraction step of extracting online input features from the online input data, an online input feature storage step of storing the online input features extracted in the online input feature extraction step, and an image stored in the image storage step. An image retrieval characterized by executing a feature retrieval step of retrieving image information corresponding to online input data by determining similarity between the line figure feature and the online input feature stored in the online input feature storage step. Law. 2. A line figure feature extraction unit for extracting line figure features that characterize the image from image information, and image storage for storing image information and the line figure features extracted by the line figure feature extraction unit in association with each other. Unit, an online input feature extraction unit that extracts online input features from online input data, an online input feature storage unit that stores the online input features extracted by the online input feature extraction unit, and an image storage unit that stores the online input features. And a feature search unit for searching image information corresponding to online input data by determining similarity between the line figure feature and the online input feature stored in the online input feature storage unit. Image retrieval device. 3. The feature search unit outputs an arbitrary number of image information corresponding to the online input data as a search result, and then combines the online input data continuously input with the online input data input so far. The image search apparatus according to claim 2, wherein the image search apparatus is configured to continue the search.