JPH087750B2 - Image retrieval device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image search apparatus that stores and manages image data as a database for searching.

[Conventional technology]

Figure 8 shows, for example, the 34th National Conference of the Information Processing Society of Japan (the first half of 1987), 3E-6, "High Definition Image File System-
FIG. 6 is a block configuration diagram showing a conventional image search apparatus as seen in the system configuration diagram of “Outline-”. In the figure, 1 is an image data input unit for inputting image data by a scanner or the like, 2 is an image data memory for storing the input image data, and 3 is an image for the image data input from the image data input unit 1. Image collation data input means for inputting collation data necessary for later retrieval, such as the name, from a keyboard or the like, 4 is an image collation data memory for storing the input collation data, and 5 is a keyboard or mouse for retrieval Search condition input means for inputting conditions, 6 is image data reading means for referring to stored collation data to obtain image data satisfying the search conditions, and 7 is image data reading means 6 from the image data memory 2 to a CRT display or the like. The image display device displays the image data obtained by

In such a conventional image search apparatus, an image name or keyword is input in advance as matching data as shown in FIG. 9, and the image name or keyword is specified as a search condition later to search for an image. Was.

[Problems to be Solved by the Invention]

The conventional image search device is configured as described above and uses the name and keyword of the image as a search condition. However, the search is performed by specifying the image feature such as the state of each figure included in the image and the connection relationship between the figures. It was difficult to enter the conditions. In order to realize the search by designating the image feature, the image feature is registered in advance in the keyword etc.,
Even if an image is obtained by designating this, the amount of data to be searched for increases, so that there is a problem in terms of function and performance such as a large amount of time required for searching image data.

The present invention has been made to solve the above problems, and an object of the present invention is to obtain an image search device capable of performing a search at a high speed by designating image features.

[Means for solving the problem]

An image retrieval apparatus according to the present invention includes an image data memory for storing image data, an image collation data memory for storing collation data for retrieving the image data stored in the image data memory, and an image data retrieval. According to similar criteria as conditions, the characteristic data of the image is extracted from the image data of the image data memory and the collation data of the image collation data memory, and described hierarchically based on the connection relationship of the devices forming the image data. Generated unique symbol and generated unique symbol data from each generated unique symbol, and the generated unique symbol data is centered on a specific device and the order of the devices directly connected to that device is fixed according to a certain rule. An image feature that sorts and stores the order of the unique symbols that make up this unique symbol data according to a certain rule. The image feature conversion accumulation and collation means includes unique symbol data generated based on image data to be searched and unique symbol data generated based on image data to be a search key. And the image data satisfying the search condition is searched for.

[Operation] The image search device according to the present invention converts the image data to be searched into unique symbol data satisfying the similar criterion as the search condition, and hierarchizes and rearranges and stores the unique symbol data. The image data to be obtained is converted into unique symbol data which also satisfies the search condition, and the unique symbol data are compared with each other to search for the desired image data at high speed.

Example of Invention

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

FIG. 1 is a block diagram showing the image retrieval apparatus of the embodiment. The same reference numerals as those in the conventional example of FIG. 8 indicate the same or corresponding parts, and the description thereof will be omitted. In FIG. 1, 8
According to the search condition input by the search condition input means 5, the feature data of the image is extracted from the image data stored in the image data memory 2 and the matching data stored in the image matching data memory 4 into unique symbol data. It is an image feature conversion accumulation / collation means for converting, accumulating and rearranging, and using this to obtain unique symbol data corresponding to image data satisfying the search condition.

 Next, the operation of the above embodiment will be described.

First, input of image data as a database is performed in the following procedure. 2 (a) and 2 (b) are system diagrams as an example of an input image. These images are input from the image data input means 1 and stored in the image data memory 2. At this time, the file names “KE1” and “KE2” are added to the stored image data of the system diagrams F1 and F2. Next, collation data is inputted from the image collation data input means 3 so that these image data can be retrieved later, and each collation data inputted is stored in the image collation data memory 4. Third
The figure shows the contents of this collation data.

FIG. 4 is a block diagram showing an example of the image feature conversion storage / collation means 8 in the image search apparatus of FIG. The image feature conversion storage / collation means 8 is used in the image data memory 2
Graphic feature data conversion means 9 for extracting graphic feature data from the image data stored in the image data and the collation data stored in the image collation data memory 4 and converting it into unique symbol data.
And a graphic feature data memory 10 for storing and accumulating the obtained unique symbol data, a rearrangement means 11 for rearranging the accumulated unique symbol data according to the search condition obtained from the search condition input means 5, and Matching means 12 for obtaining from the graphic feature data memory 10 unique symbol data corresponding to image data satisfying the search key in accordance with the search condition.

 Next, conversion / accumulation of image data is performed by the following procedure.

In advance, the search criteria input means 5 inputs similar criteria as search criteria. FIG. 5 shows the contents of this similar criterion. According to this similarity criterion, the graphic feature data conversion means 9 is calculated from the image data stored in the image data memory 2 and the matching data stored in the image matching data memory 4.
With, the feature data of the image is extracted and converted into a unique symbol,
Further, it is converted as unique symbol data for storage. The obtained unique symbol data is stored in the graphic feature data memory 10.

FIGS. 6 (a) and 6 (b) are examples of unique symbols corresponding to FIGS. 2 (a) and 2 (b) when “type of piping” in FIG. 5 is “main piping only” as a similar criterion. Yes, FIGS. 7A to 7F are examples of unique symbol data obtained from FIGS. 2A and 2B.

The unique symbol data stored in the graphic feature data memory 10 is rearranged by the rearranging means 11 in accordance with the similarity criterion as the search condition input by the search condition inputting means 5 in advance. This rearrangement not only rearranges the order of the unique symbol data, but also rearranges so that the entire unique symbol data forms a hierarchical structure according to the search condition. Sorting means 11
According to the similar criteria as the search condition input by the search condition input means 5, the rearrangement is performed so that the unique symbol data can be efficiently searched.

Next, the input, conversion, and storage of image data, which is a search key, are performed in the following procedure. The image serving as the search key is a systematic diagram similar to FIG. This image is input from the image data input means 1 and stored in the image data memory 2. The collation data of this image data is input by the image collation data input means 3 and stored in the image collation data memory 4.

The image data stored in the image data memory 2 and the collation data stored in the image collation data memory 4 according to the similar criteria as the retrieval conditions similar to those in FIG. 5 input by the retrieval condition input means 5 in advance. From the above, the graphic feature data conversion means 9 extracts the feature data of the image and converts it into a unique symbol similar to that shown in FIG. 6, and further converts it into unique symbol data similar to that shown in FIG. 7 for storage. The obtained unique symbol data is stored in the graphic feature data memory 10 as the unique symbol data of the search key.

 Next, the image data is searched in the following procedure.

The matching means 12 obtains, from the unique symbol data stored in the graphic feature data memory 10, one that matches the unique symbol data of the search key obtained in the previous procedure.
The image data reading means 6 has uniqueness data having matching data stored in the image matching data memory 4, which satisfies the search condition such as a keyword inputted by the search condition inputting means 5 in advance from the obtained unique symbol data. The symbol data is obtained, and the image data corresponding to this unique symbol data is obtained from the image data memory 2. The obtained image data is displayed on the image display device 7.

As described above, according to the above-described embodiment, the feature data of the image data is extracted by the graphic feature data conversion unit 9 provided in the image feature conversion storage / collation unit 8, and the image feature conversion storage / collation unit 8 is also extracted. Since the feature data is rearranged by the rearrangement means 11 provided in the image retrieving device, an image retrieving apparatus capable of performing a high-speed retrieval in which the feature of the image data is designated is obtained.

In the above embodiment, the image data to be the search key is inputted from the image data input means 1, stored in the image data memory 2, and the collation data of this image data is inputted from the image collation data input means 3. This is stored in the image collation data memory 4, converted into unique symbol data by the graphic feature data conversion means 9 in accordance with the search condition inputted in advance by the search condition input means 5, and the graphic feature data memory 10
Although the example of storing the unique symbol data is shown in Fig. 10, the unique symbol data is created manually or by other means from the image data as the search key according to the search condition, and the unique symbol data is stored in the graphic feature data memory
May be stored in.

Further, in the above-mentioned embodiment, the graphic feature data conversion means 9 searches the image data stored in the image data memory 2 and the matching data stored in the image matching data memory 4 by the search condition input means 5. An example of converting into a unique symbol according to the conditions and further converting into unique symbol data for storage has been shown. However, the graphic feature data converting means 9 is divided and the image data stored in the image data memory 2 and
Means for converting the collation data stored in the image collation data memory 4 into a unique symbol according to the search condition input by the search condition input means 5 and means for storing the obtained unique symbol are stored. A means for converting the unique symbol into unique symbol data may be used.

In the above embodiment, the graphic feature data conversion means 9 is
The image data stored in the image data memory 2 and the matching data stored in the image matching data memory 4 are converted into unique symbols according to the search condition input by the search condition input means 5, and further stored. For this reason, an example in which conversion is performed as unique symbol data is shown, but the same may be realized by manual work or another existing device.

In the above embodiment, the image data is input from the image data inputting means 1 and the matching data is input from the image matching data inputting means 3. However, these inputting means are not used and other existing devices are used. Image data or matching data may be obtained by using the image data, the image data may be stored in the image data memory 2, and the matching data may be stored in the image matching data memory 4.

Here, the data input using the other existing device may be only image data, only matching data, or both image data and matching data.

Further, although the case of the image search apparatus has been described in the above-described embodiment, the image data to be handled may be graphic data, pattern data, or the like, and the same effect as that in the above-described embodiment is obtained.

〔The invention's effect〕

As described above, according to the present invention, an image data memory for storing image data, an image collation data memory for storing collation data for searching the image data stored in the image data memory, and an image data A unique symbol described in hierarchization based on the connection relationship of the devices that compose the image data by extracting the characteristic data of the image from the image data in the image data memory and the matching data in the image matching data memory according to the search conditions. The generated unique symbol data is generated from each generated unique symbol, and the generated unique symbol data is rearranged around a specific device according to a certain rule in the order of the devices directly connected to the device. And an image feature conversion accumulating and collating means for rearranging and storing the order of the unique symbols constituting this unique symbol data according to a certain rule, The image feature conversion storage / collation means compares the unique symbol data generated based on the image data to be searched with the unique symbol data generated based on the search target image data to satisfy the search condition. Since the image data is searched, the image data similar to the image data can be searched at high speed, and thus the search specifying the image feature can be performed at high speed.

[Brief description of drawings]

FIG. 1 is a block configuration diagram showing an image retrieval apparatus according to an embodiment of the present invention, and FIGS. 2 (a) and 2 (b) are diagrams showing an example of a system diagram input as a database to the image retrieval apparatus. FIG. 4 is a diagram showing the contents of collation data input from the image collation data input means, FIG. 4 is a block configuration diagram showing an example of the image feature conversion storage collation means in the image retrieval apparatus of FIG. 1, and FIG. FIG. 6 is a diagram showing an example of the contents of the similar criteria input from the search condition input means, and FIGS. 6 (a) and 6 (b) are obtained by converting FIGS. 2 (a) and 2 (b) by the graphic feature data conversion means, respectively. FIG. 7 (a) to FIG. 7 (f) are diagrams showing the contents of the unique symbol data stored in the graphic feature data memory, and FIG. 8 is a diagram showing a conventional image retrieval apparatus. FIG. 9 is a block diagram showing a conventional image collation data input means. Is a diagram showing the contents of reference data is et input. 2 ... Image data memory, 4 ... Image collation data memory, 5 ... Search condition input means, 6 ... Image data reading means, 8 ... Image feature conversion storage / collation means.

Claims (1)

[Claims] 1. An image data memory for storing image data, an image collation data memory for storing collation data for retrieving the image data stored in the image data memory, and the image according to the retrieval condition of the image data. The characteristic data of the image is extracted from the image data in the data memory and the collation data in the image collation data memory, and the unique symbol described in a hierarchical manner is generated based on the connection relationship of the devices that compose the image data, and is generated. Generates unique symbol data from each unique symbol, and sorts the generated unique symbol data in the order of the equipment directly connected to the particular equipment according to a certain rule. Image feature conversion accumulation and collation means for arranging and storing the order of the unique symbols forming the symbol data according to a certain rule. The characteristic conversion storage collating means compares the unique symbol data generated based on the image data to be searched with the unique symbol data generated based on the search target image data, and finds image data satisfying the search condition. An image search device characterized by searching.