JPH0442365A - Picture retrieving device - Google Patents

Abstract

PURPOSE:To make correspondence to various retrieval conditions by a relation table in a small size by registering relation to be an element to the relation table. CONSTITUTION:This device is composed of a picture input means 1, picture storing means 2, relation exampling means 3, relation extracting and storing means 4, retrieval condition input means 5, retrieval process deciding means 6, picture reading means 7 and retrieved result output means 8. Then, the relation exampling means 3 inputs the pair of pictures or picture constituting elements exampled by a user, the relation extracting and storing means 4 extracts the relation and stores it as the description of the relation based on the inputted pair of the pictures or the picture constituting elements, and the retrieval process deciding means 6 decides a retrieval process based on the retrieval condition and the description of the relation. Thus, the size of the relation table can be reduced.

Description

[Detailed Description of the Invention] [Field of Industrial Application] This invention is an image search method that uses a computer system to manually accumulate image information such as documents and drawings, and retrieves appropriate images according to requests. It is related to the device.

[Prior art] Figure 5 shows 1, for example, Information Processing Society journal "Information Processing Press" Vol.
, 28. No. 6 (June 1987), pp, 740-
FIG. 42 is a block configuration diagram showing a conventional image retrieval device based on the configuration shown in FIG.

In the figure, ■ is an image input means for inputting images to be stored using a scanner or the like, 2 is an image storage means for storing input images in a magnetic disk device, etc., and 5 is the type of relationship specified by the user and the image or image. Search condition input means for inputting constituent elements as search conditions using a keyboard or pointing device, 7 an image reading means for reading out images that match the search conditions from the image storage means 2 based on the input search conditions, 8 this image reading means; It is a search result output means for outputting the image read out by the image reading means 7 to a graphic display or the like.

Next, the operation will be explained.

For example, images to be stored as shown in FIG. 6 (fal to fd) are inputted by the image input means 1 and stored in the image storage means 2 as image data. Image storage means 2
extracts the relationships between each image or image component based on this image data, creates a relationship table as shown in FIG. 7, and stores it together with the image data.

On the other hand, it is assumed that the user has a search request such as, for example, ``I want to see a front view of what is above the component ■ of image (a).'' In order to specify the relationship type r3 and image+8+ component ■ as search conditions in the search condition manual means 5, the user selects, for example, a group of relationship names displayed in a menu format on the graphic display 2. Select r3 from among them, and use the pointing device to point to the part where component ■ is displayed in the image (al) also displayed on the graphic display.

The image reading means 7 reads image data matching the conditions from the image storage means 2 based on the type of relationship and the type of image or image component inputted by the search condition input means 5. Here, the image storage means 2 stores a relation table as shown in FIG. The 6th and 7th lines are extracted as lines that match ``. As a result, the image C1 shown in FIG. 6 is read out. The read image data is displayed on a graphic display or the like by the search result output means 8.

[Problems to be Solved by the Invention] Since the conventional image search device is configured as described above, all possible associations as search conditions must be registered in the association table, and the size of the association table is limited. There is a problem in that the number of tables becomes enormous, and in order to change the table so that it can be searched based on unregistered relationships, it is necessary to recreate the relationship table.

This invention was made to solve the above-mentioned problems. By simply registering the elemental relationships in the association table, the small-sized association table can respond to a variety of search conditions. It is an object of the present invention to provide an image retrieval device that allows a user to register only the relationships required by the user at any time.

[Means for Solving the Problems] An image search device according to the present invention includes a relationship illustrating means for inputting an example of a pair of images or image components that satisfy the relationship according to a user's instruction, and a relationship illustrating means for inputting an example of a pair of images or image components that satisfy the relationship. a relationship extraction/storage means for extracting a relationship based on a pair of images or image components input by the user and storing the relationship as a relationship description; A search procedure determining means for determining a search procedure based on the search procedure is provided.

[Operation] In the present invention, the relationship illustrating means inputs a pair of images or image components exemplified by the user, and the relationship extraction/storage means calculates the relationship based on the input image or image component pair. is extracted and stored as a relevance description, and the search procedure determining means determines a search procedure based on the search conditions and the relevance description.

[Example] An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention, and in the figure, 1.2, 5, and 7.8 are the same or equivalent parts as in the conventional device. Reference numeral 3 indicates a relation illustrating means for inputting an example of a pair of images or image constituent elements that satisfy the relation according to a user's instruction; a relationship extraction/storage means for extracting relationships and storing them as relationship descriptions;
Reference numeral 6 denotes a search procedure determining means that determines a search procedure based on this relevance description and the search conditions input by the search condition input means 5.

Next, the operation of the above embodiment will be explained as shown in FIGS. 6 and 2 to 4.
This will be explained with reference to the figures.

The operation of the image input means 11 and the image storage means 2 is almost the same as that of the conventional device. For example, an image as shown in FIG. 6 is inputted, and a related table as shown in FIG. 2 is created based on this image. Accumulate with.

Here, the relationships to be registered in the relationship table are different from the conventional example, and only the basic relationships are sufficient.

On the other hand, it is assumed that the user has a search request such as ``I want to see the front view of something on top of the image (al component ■'').The related table contains data corresponding to this relationship. Therefore, the user first indicates this association by example using the association illustrating means 3.This association instruction is performed by, for example, displaying an image on a graphic display as shown in FIG. This is done by pointing to the main seal as an example of a pair of images or image components that satisfy the desired relationship using a pointing device or the like. Based on the pair of elements, a relationship description as shown in FIG. 4 is created and stored.

This relationship description is created by comparing pairs of images or image components that satisfy the relationship with the relationship table shown in FIG.
If it is found that a desired relationship can be achieved by combining a plurality of relationships, these combinations are registered as a relationship description. In this example, relevance r1 and relevance r2
If you combine the image (al component ■ and the image ('
Since we can see that the relationship between b) is satisfied, we can define the combination of these relationships r 10 r 2 as a new relationship r
Register as 3.

Next, in the search condition input means 5, the user inputs, for example, the component ■ of the image fal and the relevance r3 as search conditions.
instruct. In response to this instruction, the search procedure determining means 6 determines a search procedure by referring to the relevance description stored in the relevance extraction/storage means 4. In this example, the indicated relevance r3 means r 10 r 2, so this data rl.

r2 is passed to the image reading means 7.

Based on this data r10r2, the image reading means 7 first determines the relationship r between the image (al) and the component ■.
The components ■ and ■ having 1 are extracted, and the image (bl and image (C1) having the relationship r2 between the components ■ and ■ are extracted.

The search result output means 8 displays these images on a graphic display or the like.

[Effects of the Invention] As described above, according to the present invention, there is a relationship illustrating means for inputting an example of a pair of images or image components that satisfy the relationship according to a user's instruction; a relationship extraction/storage means for extracting a relationship based on a pair of images or image components that have been searched, and storing the relationship as a relationship description; Since a search procedure determining means for determining a search procedure is provided, the size of the association table can be reduced, and the user can add necessary associations at any time.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an image search device showing an embodiment of the present invention, FIG. 2 is a diagram showing an embodiment of a relation table, and FIG. 3 is a diagram showing an embodiment of a relation illustrating means. FIG. 4 is a diagram showing an example of a relationship description, FIG. 5 is a block diagram of a conventional image search device, FIG. 6 is a diagram showing an example of images to be stored, and FIG. 7 is a diagram of a conventional association table. It is a figure which shows an example. (2) is an image input means, 2 is an image storage means, 3 is a relevance example means, 4 is a relevance extraction/storage means, 5 is a search condition input means, 6 is a search procedure determining means, 7 is an image reading means, 8
is a search result output means. In addition, in the figures, the same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] An image search device that searches for an image that has a specified relationship with a specified image or a component of the image from among a plurality of images stored in advance, comprising: an image storage means for inputting an example of a pair of images or image components that satisfy the relationship according to an instruction from a user; a relationship extraction/storage means for extracting relationships based on the information and storing them as relationship descriptions; a search condition input means for inputting the type of relationship specified by the user and an image or image component as a search condition; a search procedure determining means for determining a search procedure based on the search conditions input by the search condition input means and the relevance description stored in the relevance extraction/storage means;
The image reading means reads out images matching the search conditions from the image storage means based on the search procedure determined by the search procedure determining means, and the search result output means outputs the images read out by the image reading means. An image search device characterized by: