JP2003022445A - Device and method for information retrieval, computer program for implementing the same method, and medium where the same program is recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and a method for information retrieval, which extract various feature quantities by automatically dividing various three- dimensional models into a plurality of areas, efficiently manage and store them, shorten the computation time needed to compute the similarity of a three- dimensional model having a specific area of interest, and present attribute information related to the three-dimensional model to a user, and to provide a computer program for implementing the method and a medium where the program is recorded. SOLUTION: The information retrieval device comprises an object information input means 101, an area dividing means 102, a feature quantity extracting means 103, an object information management and storage means 104, a retrieval condition input means 105, a retrieval parameter generating means 106, an object information retrieving means 107, and a retrieval result presenting means 108. User is able to perform retrieval while paying attention to a certain part in the whole three-dimensional model.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval device, an information retrieval method, a computer program for implementing this method, and a medium recording this program. More specifically, it relates to three-dimensional model data of an object. The present invention relates to an information retrieval device for presenting information about an object to a user as a key, an information retrieval method, a computer program for implementing this method, and a medium recording this program.

[0002]

2. Description of the Related Art In the search for similar objects, the similarity between a given search key object and an object in the database is determined by the similarity between the vector of feature quantities extracted from the key object and the individual objects in the database. There is a method to judge.

Generally, in a similar object retrieval apparatus for retrieving and displaying an object similar to a reference object designated by a user, some characteristic amounts of the object are calculated, and the characteristic amounts of the objects are close to each other. The similarity is calculated by calculating the distance (distance) or treating it as one multidimensional vector space that summarizes all the feature amounts without dividing the calculation for each type of feature amount and calculating the distance there. , And rank the similarity.

At this time, various types of feature quantities used for digitizing the object include, for example, color tone, lightness, saturation, and direction of lines in the image if the object is an image. ing. The distance is calculated as a Euclidean distance between the feature amounts represented by a vector for each type of feature amount.

When performing a similarity search on a three-dimensional model, it is possible to present information on an object similar to the three-dimensional model or the reference object designated by the user by using the above-mentioned similar object search method. In this method, some feature is extracted from a plurality of objects in advance and stored in a database in association with information about the object, and the user performs a similarity search based on the feature calculated from an object in the real world.

As a conventional technique for retrieving a three-dimensional model having a similar shape, for example, Takeshi Nishimura et al.
There is "Examination of similar object retrieval in three-dimensional object database" (Science Technical Report, DE2000-84, pp.103-110, 2000). This is because, in a three-dimensional model expressed using polygons, the coordinates and vector information of the polygons and the vertices that form the polygons are extracted as multidimensional vectors as physical shape features, and based on the similarity of the features. A shape similarity search is performed.

[0007]

In the similar search of a three-dimensional object, it is necessary to realize a similar search according to a searcher's intention, such as a search based on an overview of an entire body, an inspection focusing on a detailed portion, and the like. In particular, it is important to realize a part-focused search such as a search for an object in which a specific part of a certain object is similar.

In the conventional three-dimensional object similarity search method, the features of the entire model are extracted. In order to perform a search based on the features of a specific area, the model is divided in advance, and the divided model is divided into the entire model. It is necessary to register it in the database. However, with this method, management of the divided models becomes complicated, and it is difficult to unify how to divide the regions, and it is difficult to construct a practical search system.

Further, in a search focusing on a part, it is important that the part focused by the user is similar at a specific position of the three-dimensional object, and that it is similar at an arbitrary position. There are cases where it does not matter, and the fact that it is necessary to implement searches for both of these is also an obstacle to system construction. Furthermore, in order to realize a system that allows the user to perform an intuitive search, it is necessary to provide a mechanism for the user to search by focusing on a part of the overall overview of the model.

The present invention has been made in view of the above circumstances, and an object of the present invention is to solve the above-mentioned problems in the prior art and to divide a three-dimensional model into a plurality of regions, By extracting the feature amount for a region, object information retrieval based on the feature for an arbitrary region of a certain 3D model is realized, and various 3D models are automatically divided into a plurality of regions. Focusing on specific areas by extracting various feature quantities, managing and storing them efficiently 3
By reducing the calculation time required to calculate the similarity of the dimensional model,
An object of the present invention is to provide an information retrieval device, an information retrieval method, a computer program for implementing this method, and a medium recording this program, which makes it possible to present the user with attribute information associated with a three-dimensional model. .

[0011]

In order to achieve the above object, an information retrieval apparatus according to the present invention has an object information input means for reading in three-dimensional model data and attribute information of an object, and an object information input means for reading the object information input means. A region dividing unit that divides each three-dimensional model data into regions, a feature amount extracting unit that extracts a plurality of types of feature amounts from each region divided by the region dividing unit, each region divided by the region dividing unit, and Object information managing / storing means for storing the feature quantity obtained by the feature quantity extracting means in association with attribute information for each object, a search condition input means for inputting a search key object and a search area, and the search condition input means A search parameter generating means for generating a search parameter from the search key object and the search area obtained by Object information retrieval means for retrieving the desired object information by retrieving the object information stored in the object information management / storage means using the retrieved retrieval parameters, and the information obtained by this object information retrieval means to the user. It is characterized by having a search result presenting means for presenting.

Here, it is preferable that the area dividing means has a function of dividing the three-dimensional model data into predetermined areas or into arbitrary areas for each model. Further, it is preferable that each of the three-dimensional model data has a function of dividing the area division method and the size of the divided area into a plurality of types.

Further, it is preferable that the feature amount extraction means has a function of extracting a plurality of types of feature amounts for each region, and the search condition input means is a three-dimensional object serving as a search key input by a user. It is preferable to have a function of accepting the selection information of (1) and the selection information of which region of the object to focus on for the search.

On the other hand, the information retrieval method according to the present invention comprises a first step of reading the three-dimensional model data and attribute information of the object, a second step of dividing each read three-dimensional model data into areas, and division. A third step of extracting a plurality of types of feature amounts from the respective regions, and a fourth step of storing the divided regions and the feature amounts obtained from the respective regions in association with attribute information for each object. A fifth step of accepting the search key object and the search area input by the user, and a sixth step of generating search parameters from the input search key object and the search area,
A seventh step of searching the stored object information by using the generated search parameter to search for desired object information.

Here, it is preferable that the second step is to divide the read three-dimensional model data into predetermined regions or into arbitrary regions for each model. Further, it is preferable that each of the three-dimensional model data is divided into a plurality of types as to the area division method and the size of the divided area.

Further, it is preferable that the third step is to extract a plurality of types of feature quantities for each area. Further, it is preferable that the fifth step accepts selection information of a three-dimensional object, which is a search key input by the user, and selection information of which area of the object to focus on for the search.

Further, the information retrieval method according to the present invention is useful as a program for causing a computer to execute the information retrieval method, and the information retrieval method may be directly transferred to the computer or may be recorded in some recording medium and loaded. Can be executed by a computer. The present invention claims the right to use in such a form.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below in detail with reference to the preferred embodiments shown in the drawings.

FIG. 1 is a block diagram showing a schematic configuration of an information search device according to an embodiment of the present invention. The information search apparatus according to the present embodiment includes an object information input unit 101, a region division unit 102, a feature amount extraction unit 103, an object information management / storage unit 104, a search condition input unit 105, a search parameter generation unit 106, an object information search. It comprises means 107 and search result presentation means 108.

Object information input means 10 of the object information storage device
1 acquires the attribute information of an object and the three-dimensional model data of this object. The area dividing means 102 is the image input means 1
The three-dimensional model data acquired by 01 is divided into a plurality of areas. The feature amount extracting means 103 is the area dividing means 10
Region information, which is the feature amount calculated for each of the two divided regions, is extracted.

The object information managing / storing means 104 associates attribute information and area information relating to each three-dimensional object with each other and stores them in a database. Search condition input means 10
Reference numeral 5 receives a key object name and a search condition used in the search from the user. Search parameter generation means 106
Generates a search sentence based on the key object name and the search condition input by the search condition input means 105.

The object information searching means 107 acquires the object information by inputting the search sentence generated by the search parameter generating means 106. The search result output unit 108 finally presents the object information acquired by the object information search unit 107 to the user.

Next, the operation of the information search apparatus according to this embodiment will be described. FIG. 2 is a flowchart of processing for constructing an object database in the information search device according to this embodiment.

In the information retrieving apparatus according to this embodiment, three-dimensional model data of an object to be retrieved is
It is generated using a three-dimensional digitizer or CAD software. That is, in step 11, the three-dimensional model data generated by the above-mentioned three-dimensional digitizer or CAD software by the object information input means 101 and the object of retrieval, for example, the object name or an image of the object photographed, etc. Object information consisting of the attribute information of is input.

In step 12, the object information input means 10
From the three-dimensional model data input at 1, the area dividing means 102 performs smoothing, size normalization or polygon increase / decrease as needed, and then performs boundary extraction, or according to a predetermined section. , Split into multiple regions. In step 13, the area dividing means 1
From the plurality of regions divided by 02, the feature amount extraction means 103 respectively calculates, for example, the distribution of the vertex coordinates of the model data and the distribution of the direction of the normal vector of the polygon.
A plurality of feature quantities are extracted.

In step 14, the object information managing / storing means 104 determines whether or not the model being processed is the first object based on the object name to be searched input in step 11. The object database stored in the object information managing / storing means 104 is searched and checked. In step 14, if the input object name already exists, the process ends.

In step 14, if the input object name does not exist and it is the first object name, step 15
Then, by the object information management / storage means 104, step 1
An object object is created from the object name and object information of the object input in step 1, and in step 16, this is associated with the feature amounts corresponding to the plurality of regions extracted in step 13 to create an object as a child object. Store in database.

The above process is repeated for a plurality of objects to build an object database. The structure is shown in FIG. In FIG. 3, 20 is a physical object, 21, 22,
... 2n indicates a child object. In the following, a flow of search processing in the information search device according to the present embodiment using the physical object and the child physical object will be described.

The user of the information search device inputs a search key such as an object name by using the search condition input means 105. Further, the search condition input means 105 specifies an area to be searched, and if necessary, weights features or viewpoints used for the search, or item names such as attribute information or bibliographic information desired as a search result. Enter search parameters such as.

The search parameter generating means 106 creates a query sentence for the database using a query language such as SQL sentence based on the search key, the target area and the search parameter. The object information search means 107 uses the inquiry sentence generated by the search parameter generation means 106 to perform a similarity search of the object database and obtain a result. The search result output unit 108 finally presents the object information and the like to the user based on the search result acquired by the object information search unit 107.

Example 1 The object information creator (hereinafter referred to as the creator) first wants a user of this search device, such as a pottery in various parts of Japan, to "know something about it." The object is modeled by itself using a three-dimensional digitizer or the like. Also, the name and place of origin of the pottery,
Prepare the producer name or attribute information such as a photograph.

As a specific example, the flow of processing when information about the cup “wareware A” of the pottery is stored using the information retrieval apparatus according to this embodiment will be described below.

The creator models the container A using a three-dimensional digitizer or the like, and names the object "model A" modeled by the object information input means 101,
Type on your computer. In addition, the attribute information prepared in advance is input as the object information. The object information input means 101 assigns to the input model a uniquely determined object ID such as a number string generated based on the input time (year, month, day, hour, second), for example, in NFF format. , Model data handled by the area dividing unit 102.

The area dividing means 102 receives the model data to which the object ID is assigned from the object information inputting means 101, and, for example, standardizes the circumscribed rectangular parallelepiped and then divides it into preset sections, or on the model. Divide the model data into a predetermined arbitrary number and size using a method such as dividing a set of points existing within a certain Euclidean distance or an adjacent distance from an arbitrary point into a region. .

Further, the area dividing means 102 applies to the divided models, for example, numbers given in the order of division,
Assign a region ID that is uniquely determined in the model. The feature amount extraction unit 103 receives the model divided by the region dividing unit 102, and, for example, the number of polygon vertices existing in the region, a histogram of the direction of the normal vector of the polygon, or the region is further subdivided into polygon vertices. A plurality of feature quantities representing the shape are extracted for each area by a method such as a distribution histogram.

With respect to the above-mentioned respective processes, for example, the methods described in the following documents can be effectively used. A Study of Similar Object Search in 3D Object Database by Tsuyoshi Nishimura et al. (Science Tech., DE2000-84, pp.103-110,
2000).

The object information managing / storing means 104 manages an object-oriented database, and handles a data structure as shown in FIG. 3, for example. The object information managing / storing means 104 first checks whether or not the object information including the object name “container A” as shown in FIG. 4A is already stored in the database.

If the object name "Unit A" does not exist in the database, the object information management / storing means 104 newly creates an object class object (20 in FIG. 3), and the object name "Unit A". After storing the other object information in this object, a new object child object area object Domain and divided model feature value Domain Fe
ature (21, 22, ..., 2n in FIG. 3) is created, and the extracted feature quantity of the feature quantity extraction means 103 and the area ID are stored in association with each other.

On the other hand, when the object name “container A” already exists in the database, the object information management / storage means 104.
Notifies the creator that the "device A" exists in the database, and ends the storage process. Similarly, for ceramics other than "ware A", the process of storing the feature amount is repeated to create a plurality of object objects and construct an object database corresponding to a multimedia picture book.

In the following, as a specific example, the user searches for an example of a pottery stored in the object database, obtains pottery similar to the selected pottery, and information on the pottery, and retrieves the information from the information retrieval device. The flow of processing when used is shown. Here, it is assumed that the information on the vessels A, B, C, and D is stored in the object database by repeating the process of storing the feature quantity of the information retrieval apparatus in advance.

As a first example, an example is shown in which the three-dimensional model data is divided into predetermined areas to define the areas. For the sake of explanation, as shown in FIG. 4, it is assumed that a model projected on a two-dimensional plane is used and the features are extracted for each region when the circumscribed rectangular parallelepiped is divided at equal intervals. At this time, regarding each model, the same area ID is assigned to the areas at the same position among the areas obtained by dividing the circumscribed rectangular parallelepiped into the same number.

FIGS. 5A to 5D are enlarged views of the region 4 in the regions 1 to 18 of each object in FIG. The search condition input means 105 refers to the object information stored in the object database, for example, and randomly displays on the screen a photograph of the ceramics, a model overview of the ceramics, and a region selection screen. When the user selects an arbitrary area from the overview of this model and the area selection screen, the object ID of the ceramic and the selected area ID are acquired.
Here, it is assumed that the user has selected the area 4 of the container A.

The search parameter generation means 106 narrows down the object area object Domain of each object by the area ID based on the object ID and the area ID of the search key acquired by the search condition input means 105, and then the Domain Feature From the object area object stored in the object area object, search for an object area object in which the feature quantity of each area has a degree of similarity with the feature quantity related to the area specified by the search key, and search the object from the parent object of the corresponding object. Generate a database query statement such as getting the name.

In this example, based on the similarity of the features of the area 4, an object having a high similarity of a certain value or more is searched for in a specific area on the model, and the object object that is the parent object of the area is searched. A query statement for returning the name of the object and the image file name of the object is generated.

The object information searching means 107 searches the object database for similarity using the inquiry sentence generated by the search parameter generating means 106, and obtains desired information. In this example, the similarity of the container C (of the region 4) is increased based on the characteristics of the region 4 and is acquired as the search result. Based on the search result, the search result display unit 108 presents the user with the names and images of the ceramics and the container C that are found to have a high degree of similarity.

According to the information retrieval apparatus of this embodiment,
As described above, it is possible to search for an object having a similar shape in the area designated by the user.
In the above example, the area 4 is set as a search target with the container A as a key, but the same applies when another area is designated in another object. That is, according to the information search device of the present embodiment, it is possible to realize a fine-tuned search focusing on a part.

Further, in the above embodiment, as a means for dividing the area, a method of dividing the circumscribed rectangular parallelepiped into equal intervals is shown as an example, but the division method is not limited to this, and the predetermined area may be any area. The shape may be any size. Further, the regions may overlap with each other or may have a portion not included in any region. Hereinafter, a specific example of this case will be shown.

[Embodiment 2] As a second example, an example in which three-dimensional model data is divided into arbitrary regions for each model will be described. Also here, as in the first example, by repeating the process of storing the feature amount of the information search device in advance, for example, the information of the device A, the device B, the device C, and the device D is stored in the object database. Be present. Further, for the sake of explanation, a model projected on a two-dimensional plane is used, and as shown in FIG. 6, it exists within a certain adjacent distance from any point on the model within a certain adjacent distance. It is assumed that a set of points is divided into a plurality of areas, and features are extracted for each area.

The search condition input means 105 refers to the object information stored in the object database, for example, and randomly displays a photograph of the ceramics and a call area selection screen of the model overview of the ceramics on the screen. When the user selects an arbitrary area from the view and area selection screen of the model, the object ID of the ceramic and the selected area ID are acquired. In this example, it is assumed that the user has selected the area 2 of the container A.

On the basis of the object ID and the area ID of the search key acquired by the search condition inputting means 105, the search parameter generating means 106, for example, with respect to all object area objects, has a constant feature amount and search key extracted from each area. Search for an object region object that has a degree of similarity above the value,
Generate a database query statement such as obtaining the object name from the parent object of the corresponding object.

In this example, based on the similarity of the area 2 of the container A, all area objects on the model are searched for an object having a high similarity of a certain value or more, and the object that is the parent object of the area is searched. Returns the name of the object and the image file name of the object.

The object information searching means 107 uses the inquiry sentence generated by the search parameter generating means 106 to perform a similar search on the object database to obtain desired information. In this example, for example, as shown in FIG. 7, the similarity of the features in the region 3 of the container B becomes high and is acquired as a search result. Based on the search result, the search result display means 108 presents the user with the names and images of the ceramics and the container B that are found to have a high degree of similarity.

In the above example, the area specified by the user is
It is possible to search for an object having a similar shape at any position. It should be noted that each of the above-mentioned embodiments shows one example of the present invention, and the present invention should not be limited to these.

For example, in the above example, the area A is set as the search key with the container A as the key, but any other area may be designated as the search key, and fine-tuned search focusing on the part is realized. It is possible. Further, here, when the area is divided, the barycentric coordinates of the divided area are extracted as a feature, and the barycentric coordinates are narrowed down at the time of search, so that a specific portion is specified as in the first example. It is also possible to realize a search on the condition that it is located at.

Further, in the above example, as a method of dividing the area, a set of points existing within a certain adjacent distance from an arbitrary point within a certain adjacent distance centered on an arbitrary point on the model is defined as the area. As an example, the method of dividing as
The method of division is not limited to this, and for example, 3
In the polygons that make up the dimensional model, any method, any shape, such as centering the center of gravity of polygons with a difference between the normals of adjacent polygons being a certain value or more, or using the Euclidean distance instead of the adjacent distance It does not matter if it is divided by the size of. Further, the regions may overlap with each other or may have a portion not included in any region.

Further, in the first and second examples of the above-described embodiment, an example has been shown in which the user performs a search by designating one divided area, but a plurality of areas may be designated. Often,
Further, the degree of importance in each area may be designated.

Further, in the first and second examples of the above embodiment, a search is performed based on a feature extracted from a model obtained by dividing a predetermined region, and extraction is performed for each region divided according to the model. Although an example has been shown for each of the searches based on the above characteristics, it is also possible to combine these to extract the characteristics in both methods, and to implement a narrowed search by combining both methods.

Further, in the first and second examples of the above-described embodiment, the regions are divided into regions divided at a constant interval or within a certain distance in each method, and the regions are divided only for each one type. However, an area may be divided by an arbitrary number and an arbitrary number, and for example, a search in which a finely divided area and a largely divided area are combined can be realized.

Furthermore, in the first and second examples of the above-described embodiment, the model mapped to the two-dimensional plane is used for the sake of explanation, but the three-dimensional model is also divided into regions. Is possible. FIG. 8 shows an example in which the three-dimensional space is divided into areas.

The basic idea of information retrieval according to the present invention is not limited to actual object retrieval, but is widely used for retrieval of images, images, etc. by using the feature amount of a portion in an image. It is applicable.

The information search method according to the present invention is particularly useful when the program is coded to be executed by a computer, and the program is directly transferred to the computer via a network, or It is possible to cause a computer to execute by recording on some recording medium and loading it.

[0062]

As described above in detail, according to the present invention, the three-dimensional model data relating to the object is divided into a plurality of regions, and the feature amount relating to the region of the three-dimensional model data is divided from the plurality of regions. It is possible to efficiently store, manage, and store the object information and the features related to the area because the extracted information is associated with the object information regarding the underlying three-dimensional model data and stored together. In addition, when the user wants the information about the object, by designating the part of the object to be focused on, it is possible to realize the information search device capable of presenting the information about the object similar to the part. can get.

Further, in the information retrieval method according to the present invention, the three-dimensional model data regarding the object is divided into a plurality of regions, and the feature amount related to the region of the three-dimensional model data is extracted from the plurality of regions. Based on 3
The information related to the dimensional model data is stored together in association with the object information, which makes it possible to efficiently store, manage, and store the object information and the features related to the area, and when the user wants the information about the object, By designating the part of the object of interest, it is possible to realize the information retrieval method capable of presenting the information about the object to which the part is similar.

Further, the information search method according to the present invention is useful as a program for causing a computer to execute the information search method, and the present invention has a great effect on the usage form in such a format. Is.

[Brief description of drawings]

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

FIG. 2 is a flowchart of processing for constructing an object database in the information search device according to the embodiment.

FIG. 3 is a conceptual diagram of a physical object configuration.

FIG. 4 is a diagram illustrating an example of first area division in the information search device according to the embodiment.

5 is an enlarged view of a region 4 in FIG.

FIG. 6 is a diagram illustrating an example of second area division in the information search device according to the embodiment.

7 is an example of similar areas in FIG.

FIG. 8 is a diagram showing an example in which a three-dimensional space is divided into areas.

[Explanation of symbols]

101 Object information input means 102 area dividing means 103 feature extraction means 104 Object information management / storage means 105 Search condition input means 106 search parameter generation means 107 object information retrieval means 108 Search result presenting means

Continued front page    F term (reference) 5B050 AA03 BA13 DA10 EA06 EA18                       FA02 FA19 GA08                 5B075 NK39 PR06 QM08                 5L096 AA09 BA20 CA24 DA01 GA19                       HA08 JA11 JA16 KA09

Claims (12)

[Claims] 1. An object information input means for reading three-dimensional model data and attribute information of an object, an area dividing means for dividing each three-dimensional model data read by the object information input means into areas, and this area dividing means. Feature amount extraction means for extracting a plurality of types of feature amounts from each divided region, and each region divided by the region dividing means and the feature amount obtained by the feature amount extraction means are associated with attribute information for each object. Object information management / storage means for storing, search condition input means for inputting a search key object and a search area, and search parameter generating means for generating a search parameter from the search key object and search area obtained by the search condition input means And storing the object information management / storage means by using the search parameters obtained by the search parameter generation means. An information retrieval apparatus comprising: an object information retrieval means for retrieving desired object information by retrieving the desired object information; and a retrieval result presentation means for presenting the information obtained by the object information retrieval means to a user. . 2. The information search according to claim 1, wherein the area dividing means has a function of dividing the three-dimensional model data into predetermined areas or into arbitrary areas for each model. apparatus. 3. The area dividing means has a function of dividing the three-dimensional model data into a plurality of types of area division method and area size to be divided. The information retrieval device described in 2. 4. The information retrieving apparatus according to claim 1, wherein the feature amount extraction means has a function of extracting a plurality of types of feature amounts for each area. 5. The search condition input means has a function of accepting selection information of a three-dimensional object, which is a search key input by a user, and selection information of which area of the object to focus on when performing a search. The information search device according to claim 1, wherein 6. A first step of reading three-dimensional model data and attribute information of an object, and a second step of dividing each read three-dimensional model data into regions.
And a third step of extracting a plurality of types of feature amounts from each divided region, and storing each divided region and the feature amount obtained from each region in association with attribute information for each object. The fourth step, the fifth step of accepting the search key object and the search area input by the user, the sixth step of generating the search parameter from the input search key object and the search area, and the generated search A seventh step of searching the stored object information using a parameter to search for desired object information. 7. The second step comprises the steps of reading each 3
7. The information search method according to claim 6, wherein the dimensional model data is divided into predetermined areas or into arbitrary areas for each model. 8. The method according to claim 6, wherein the second step is to divide the area division method and the area size to be divided into a plurality of types for each of the three-dimensional model data. Or the information retrieval method described in 7. 9. The information search method according to claim 6, wherein the third step is to extract a plurality of types of feature quantities for each area. 10. The fifth step is to accept selection information of a three-dimensional object, which is a search key input by a user, and selection information of which area of the object to focus on when performing a search. 10. The method according to claim 6, wherein
Information retrieval method described in section. 11. A program for causing a computer to execute the information search method according to claim 6. 12. A recording medium in which a program for causing a computer to execute the information search method according to claim 6 is recorded.