Patent Grant
7899274
This invention relates to a technique for retrieving contents including a drawing, video, image, three-dimensional Computer Aided Design (3D-CAD) data, volume data and the like (e.g. entity that can be perceived by the sense of sight, hearing, taste, touch or smell).
In the industry, in order to develop the product having the competitiveness and early bring it to market, it is desired to effectively utilize the technical resources such as drawings of the existing products. A technique to efficiently retrieve drawings in which necessary components are drawn is required in order to find out necessary information from among a large volume of drawings, which have been accumulated up to now. Then, as a simple method, a text-based search is frequently used, in which a keyword such as a component name assigned to the drawing is input. In addition, recently, in order to realize a search using the shape of the component, which is difficult to represent it by the text, a similar shape search using shape information of the component drawn in the drawing is being used.
On the other hand, a following study is reported. Specifically, when a subject plays janken, a Magnetic Resonance Imaging (MRI) apparatus measures changes of the blood flow according to the brain activity for each second, and the computer extracts only portion relating to action commands of the brain activity from its image to analyze and interpret which action in the janken is played. The computer sends the result to a robot hand to cause it to perform the same action as the subject plays. However, the study does not mention the search for the contents, which concretely exist.
However, in the conventional similar shape search techniques, the shape information extracted from the drawing is a shape feature quantity predefined by a system developer, and according to a query drawing designated by a searcher, there is a case where any similar result expected by the searcher cannot be obtained. This is because an appropriate shape feature quantity along the intention of the searcher cannot be always obtained by a predetermined extraction method of the shape feature quantity used in the search.
In addition, although it is necessary to prepare a query drawing, there are a lot of cases where there is no appropriate query drawing in hand. In such a case, although it is considered that the searcher draws the component he or she would like to find out by a sketch input or the like, it is often difficult to draw the component the searcher images in mind as he or she wants by the sketch.
Therefore, an object of this invention is to provide a new technique to retrieve the contents.
In addition, another object of this invention is to provide a new technique to enable the retrieval of the contents the searcher images in mind.
Furthermore, still another object of this invention is to provide a new technique to enable the searcher to retrieve the contents more easily.
A search method for retrieving a content the human being can perceive according to a first aspect of this invention includes: obtaining a query brain image representing a brain activity state of a searcher when perceiving or imagining the content to be retrieved; identifying the content corresponding to the query brain image by using a structure associating a content with a brain image representing the brain activity state when perceiving the content; and outputting the content corresponding to the query brain image.
Thus, by using the query brain image, without specifically indicating the content such as drawing, the searcher can extract the pertinent content only by imaging the contents. That is, it becomes possible to easily carry out the retrieval of the content.
In addition, the aforementioned structure may be a database storing a DB brain image representing the brain activity state when perceiving the content in association with the content. In such a case, the aforementioned identifying may include: searching the database by the query brain image to identify the DB brain image having predetermined similarity with the query brain image; and identifying the content corresponding to the identified DB brain image from the database. Thus, by preparing the DB brain image in association with the content in advance, it becomes possible to identify the image similarity between the DB brain image and the query brain image.
Furthermore, the aforementioned predetermined similarity may be determined based on a difference between density of a pixel of the DB brain image and density of a corresponding pixel of the query brain image or a difference between density of a pixel of a portion of the query brain image and density of a corresponding pixel of the DB brain image.
Furthermore, the aforementioned predetermined similarity may be determined based on a distance between a feature quantity of the DB brain image and a feature quantity of the query brain image.
In addition, a user ID of a user whose DB brain image was measured may be associated with the DB brain image in the aforementioned database. In such a case, the aforementioned identifying may include narrowing the DB brain images by the user ID of the searcher in the database. This is because the brain image may be different when the searcher is different.
Furthermore, the aforementioned identifying may further include, when the DB brain image having the predetermined similarity cannot be extracted from the narrowed DB brain images, searching the database regardless of the user ID of the searcher or by using the user ID other than the user ID of the searcher.
In addition, a user ID of a user whose DB brain image was measured may be registered in association with the DB brain image in the database. Furthermore, the DB brain image may be clustered for each user ID. In addition, each cluster may include a test registration brain image representing the brain activity state of a pertinent user when perceiving a test content. Then, in such a case, the first aspect of this invention may further include: obtaining a test brain image representing the brain activity state of the searcher when perceiving the test content; and comparing the test brain image with the test registration brain image for each cluster, which is stored in the database, to identify the cluster to which the test brain image belongs. Furthermore, in such a case, the aforementioned identifying may include narrowing the DB brain images by the cluster to which the test brain image belongs in the database. When there is a predetermined tendency in the brain image, the more appropriate content can be extracted by introducing such clusters.
In addition, the first aspect of this invention may further include registering, in the database, the query brain image in association the content identified as the retrieval result for the query brain image. In addition, the first aspect of this invention may further include registering, in the database, the test brain image in association with the test content. In order to reuse the query brain image or test brain image, these steps are introduced.
Furthermore, the aforementioned content may be a single drawing. In such a case, the search method according to the first aspect of this invention may further include: identifying, by a data structure managing combinations of a drawing that is the content and a drawing including the specific drawing, a combination for a drawing that is the content identified at the identifying. Then, the aforementioned outputting may include attaching a mark to the drawing identified at the identifying to output the identified combination. When plural drawings are integrated, the pertinent portion can be presented for the searcher.
Furthermore, the aforementioned identifying may further include sorting the DB brain images having the predetermined similarity with the query brain image in a descending order based on the similarity degree of the identified DB brain images. In addition, the aforementioned outputting may include outputting the content corresponding to the query brain image in an order of the sorting. Furthermore, the search method according to the first aspect of this invention may include obtaining a second query brain image of the searcher at a timing designated by the searcher or at an arbitrary timing; and carrying out the identifying and the outputting for the second query brain image. By this processing, it becomes possible to narrow the contents much similar to the contents the searcher imagines in mind.
In addition, the aforementioned obtaining may include inputting a feature quantity of the content to be retrieved into a neural network in which the study is carried out by using a feature quantity of a content as input data and using, as output data, a brain image representing the brain activity state when perceiving the content to obtain a corresponding brain image. When the content to be retrieved or similar content is held, the retrieval can be easily carried out.
In addition, the aforementioned structure may be a neural network in which the study is carried out by using a brain image as input data and using a content corresponding to the brain image as output data.
A search apparatus for retrieving a content the human being can perceive according to a second aspect of this invention includes: an obtaining unit that obtains a query brain image representing a brain activity state of a searcher when perceiving or imagining a content to be retrieved; an identifying unit that identifies a content corresponding to the query brain image by using a structure associating a brain image representing a brain activity state when perceiving a content with the content; and an output unit that outputs the content corresponding to the query brain image.
In addition, the aforementioned structure may be a database storing the DB brain image representing the brain activity state when perceiving the content in association with the content. In such a case, the aforementioned identifying unit may search the database by the query brain image, identify the DB brain image having predetermined similarity with the query brain image, and identify the content corresponding to the identified DB brain image from the database.
Furthermore, a user ID of a user whose DB brain image was measured may be registered in association with the DB brain image in the database. Then, the DB brain images may be clustered for each user ID. Furthermore, each cluster may include a test registration brain image representing the brain activity state of a pertinent user when perceiving a test content. In such a case, the search apparatus according to the second aspect of this invention may further include a unit that identifies a content to which the most users register the DB brain image by searching the database or a predetermined content, as the test content. Furthermore, the aforementioned obtaining unit may obtain the test brain image representing the brain activity state of the searcher when perceiving the test content, and the aforementioned identifying unit may compare the test brain image with the test registration brain image for each cluster, which is stored in the database, and identify the cluster to which the test brain image belongs to narrow the DB brain images by the cluster to which the test brain image belongs in the database. This processing is carried out in order to carry out an accurate retrieval by carrying out an appropriate clustering.
Furthermore, the second aspect of this invention may further include: a unit that registers, when a plurality of contents having similarity equal to or more than a predetermined reference are detected, one brain image in association with the plurality of contents into the database. Thus, because it is necessary to measure the brain image only once for the extremely similar drawings or the like, the workload can be reduced when accumulating the data.
Furthermore, the search apparatus according to the second aspect of this invention may further include a unit that presents an arrangement of the DB brain images based on feature quantities of corresponding contents for a user; a unit that presents the contents in the same arrangement as the arrangement of the DB brain images for the user; and a unit that registers a relation between the contents and the DB brain images, which are specified by the user, into the database.
Similarly, the second aspect of this invention may further include: a unit that presents an arrangement of the DB brain images based on feature quantities of the DB brain images for a user; a unit that presents corresponding contents in the same arrangement as the arrangement of the DB brain images for the user; and a unit that registers a relation between the contents and the DB brain images, which are specified by the user, into the database. Thus, it becomes possible to register new data into the database and utilize it for the new search.
Incidentally, it is possible to create a program for causing a computer to execute the search method according to the present invention. The program is stored into a storage medium or a storage device such as, for example, a flexible disk, a CD-ROM, a magneto-optical disk, a semiconductor memory, or a hard disk. In addition, the program may be distributed as digital signals over a network in some cases. Data under processing is temporarily stored in the storage device such as a computer memory.
First, an outline of an embodiment of this invention will be explained. In this embodiment, by using the similarity of brain images (e.g. two-dimensional image, three-dimensional image, time sequence images or the like) of a searcher, which are measured by an apparatus measuring a brain activity area such as a functional Magnetic Resonance Imaging (fMRI) apparatus or optical topography apparatus, a search apparatus for retrieving the content the human being can perceive is configured. By this technique, without creating a query drawing the searcher imagines by the sketch, it becomes possible to retrieve the drawing by using the brain image when the searcher imagines the query drawing. In the following, an example of retrieving the drawing will be explained.
In this embodiment, as a preprocessing, as shown in the right side of
The brain image input unit 7 refers to the drawing DB 1 or the brain image DB 5 to identify a test drawing or the like and also carries out a processing to present it for the user. In addition, in order to register the query brain image into the brain image DB 5 in association with a drawing selected by the searcher, the input and output unit 13 outputs drawing selection instruction data by the searcher to the brain image manager 3, and outputs an instruction to register the query brain image into the brain image DB 5 to the search processor 9.
The brain image DB 5 also holds data as shown in
Next, a processing flow of the search apparatus shown in
Next, the search processor 9 identifies one unprocessed DB brain image in the brain image DB 5 (step S3). Then, the search processor 9 calculates a degree of the similarity between the query brain image and the DB brain image identified at the step S3, and stores the degree of the similarity into a storage device such as a main memory (step S5). At the step S5, for example, it is possible to calculate a difference between the density of a pixel of the query brain image and the density of a corresponding pixel of the DB brain image, and calculate, as the degree of the similarity, the total sum of the absolute values of the differences. In such a case, the less the total sum of the absolute values of the differences, the higher the degree of the similarity is. Incidentally, the differences of the pixel density may be calculated only for a portion in both of the query brain image and the DB brain image.
In addition, it is also possible that an image feature quantity of the color, shape or the like is extracted from the query brain image to generate a feature quantity vector from the image feature quantity, an image feature quantity of the color, shape, or the like is extracted from the DB brain image to generate a feature quantity vector from the image feature quantity, and the distance between these feature quantity vectors is calculated as the degree of the similarity. Also in such a case, the shorter the distance is, the higher the degree of the similarity is. Incidentally, as for the image feature quantity, a technique described in Takayuki Baba et al., “A shape-based part retrieval method for mechanical assembly drawings”, the technical study report of the Institute of Electronics, Information and Communication Engineers, PRMU2004-225, pp. 79-84 (2005) can be used.
Then, the search processor 9 judges whether or not all of the DB brain images in the brain image DB 5 have been processed (step S7) When there is an unprocessed DB brain image, the processing returns to the step S3. On the other hand, when all of the DB brain image in the brain image DB 5 have been processed, the search processor 9 sorts the drawings corresponding to the DB brain images according to the data structure as shown in
The input and output unit 13 outputs the drawings in the sorting order based on the sorting result stored in the search result storage 11 (step S11). For example, the output screen includes a “next” button and a “select” button. Then, when an inappropriate drawing was displayed, the searcher clicks the “next” button. When the “next” button is clicked, the input and output unit 13 displays the next ordered drawing. When the drawing the searcher visually checked or imagined is output, the searcher clicks the “select” button. In response to the selection of the drawing, the input and output unit 13 outputs the drawing ID of the drawing to the brain image manager 3. In addition, the input and output unit 13 instructs the search processor 9 to register the query image stored in the brain image storage 8 into the brain image DB 5, and in response to the instruction, the search processor 9 registers the query brain image into the brain image DB 5. In addition, the brain image manager 3 registers the drawing ID in association with an ID of the query brain image into a table as shown in
By carrying out the aforementioned processing, the drawing the searcher keeps in mind can be extracted without specifically drawing the drawing by himself or herself.
In the first embodiment, the narrowing of the DB brain images is not specifically indicated. In the second embodiment, because it is considered that the brain images are somewhat different according to the searcher, the user ID of the user whose DB brain image is measured is registered in association with the DB brain image. For example, the data as shown in
Then, as for the query brain image of the searcher A whose user ID is “A”, first, the comparison target is narrowed to the DB brain images to which “A” is registered as the user ID. In an example of
Thus, because the comparison is carried out along the tendency of the searcher, it becomes possible to obtain more accurate search result.
In the second embodiment, the comparison target is narrowed based on the user ID. However, because there are users having the same tendency in the brain image, it is possible to carry out a clustering based on the similarity of the brain image, not user ID.
For example, the brain image manager 3 carries out the clustering of the brain images by using the K-means method as shown in
Then, the brain image manager 3 judges whether or not either of the cluster centers moves equal to or more than a predetermined reference (step S27). When either of the cluster centers moves equal to or more than the predetermined reference, the processing returns to the step S23. On the other hand, when any of the cluster centers do not move, the brain image manager 3 registers the clusters of the DB brain images into the brain image DB 5 (step S29). The cluster is registered to the DB brain image to be processed at the step S23 according to the classification at the step S23, and the cluster is registered to the DB brain image other than the DB brain image to be processing at the step S23 based on the user ID. Namely, when the user ID of the DB brain image to which the cluster “a” was registered is “A”, the cluster “a” is registered to the DB brain image whose user ID is “A”. At the step S29, the data as shown in
A state where such data is managed in the brain image DB 5 is schematically indicated in
Although the K-means method is adopted for the clustering method, another method can be used for the clustering in this embodiment.
When such data is managed in the brain image DB 5, the DB brain images to be compared at the step S3 in
Next, the search processor 9 compares the obtained brain image with the DB brain image corresponding to the test drawing in each cluster (step S33). That is, the search processor 9 calculates the degree of the similarity between the obtained brain image and the representative DB brain image in each cluster, and stores the degree into the storage device such as the main memory. The calculation of the degree of the similarity is the same as the aforementioned one. Then, the search processor 9 identifies the cluster having the highest degree of the similarity, and limits the processing target to the DB brain images in the cluster (step S35). Namely, for example, in the table as shown in
Incidentally, by carrying out the clustering again at a predetermined interval or at an arbitrary timing, the clusters are dynamically changed by the brain image manager 3 in order to maintain and manage appropriate clusters.
Thus, because the processing target can be narrowed to the DB brain images of the users having the similar tendency, it becomes possible to obtain an appropriate search result at high speed.
As shown in the upper right of
Then, a processing as shown in
Next, the search processor 9 identifies one unprocessed DB brain image in the brain image DB 5 (step S43). Then, the search processor 9 calculates the degree of the similarity between the query brain image and the DB brain image identified at the step S43, and stores the degree into the storage device such as the main memory (step S45). At the step S45, for example, it is possible that the difference between the density of a pixel of the query brain image and the density of a corresponding pixel of the DB brain image is calculated, and the total sum of the absolute values of the differences are calculated as the degree of the similarity. Incidentally, it is also possible that the difference of the pixel density between only portions in both of the query brain image and the DB brain image is calculated.
Moreover, it is also possible that an image feature quantity of the color, shape or the like is extracted from the query brain image to generate a feature quantity vector from the image feature quantity, a image feature quantity of the color, shape or the like is extracted from the DB brain image to generate a feature quantity vector from the image feature quantity, and the distance between these feature quantity vectors is calculated as the degree of the similarity.
Then, the search processor 9 judges whether or not all of the DB brain images in the brain image DB 5 have been processed (step S47). When there is an unprocessed DB brain image, the processing returns to the step S43. On the other hand, when all of the DB brain image in the brain image DB 5 have been processed, the search processor 9 sorts the DB brain images based on the degrees of the similarity in the descending order (step S49). For example, the brain image IDs of the DB brain images are arranged based on the degree of the similarity in the descending order. After that, the search processor 9 identifies the drawing (i.e. the single-component drawing) corresponding to the brain image having the highest degree of the similarity (step S51), extracts the drawing from the drawing DB 1, and stores the drawing into the search result storage 11. Furthermore, the search processor 9 identifies the multiple-component drawing including the single-component drawing identified at the step S51 from the data structure as shown in
Here, the searcher refers to the presented multiple-component drawing 31 and the mark 41 to judge whether or not the target drawing is presented. When the target drawing is not presented, he or she instructs the brain image input unit 7 to input the query brain image again. The brain image input unit 7 accepts the instruction of the re-input of the query brain image (step S55: No route), obtains the query brain image when he or she imagines the drawing to be retrieved again after visually checking the output (i.e. presented content) and stores the query brain image into the brain image storage 8 (step S57). In
On the other hand, when it is judged that the target drawing is presented, the searcher inputs the selection instruction to the input and output unit 13. When accepting the selection instruction from the searcher (step S55: Yes route), the input and output unit 13 registers the query brain image into the brain image DB 5, and registers the brain image ID of the query brain image and the drawing ID of the corresponding drawing (i.e. simple-component drawing) into the data structure shown in
Even if such a multiple-component drawing is used, it is possible to extract necessary drawings and present them for the searcher.
Incidentally, although a processing example in which the query brain image is obtained again at the step S57 is shown, it is possible to search the drawing DB 1 for drawings similar to the identified single-component drawing, for example, to present them for the searcher.
The search processing can be variously changed, and a search processing as shown in
First, the input and output unit 13 initializes a counter i to “1” (step S61). Then, the brain image input unit 7 obtains a query brain image representing the brain activity when the searcher visually checks the drawing to be retrieved or when the searcher imagines the drawing to be retrieved, and stores the query brain image into the brain image storage 8 (step S63). The entire query brain image may be used, and only a pre-specified active area may be used. When only portion of the query brain image is used, only corresponding portion of the DB brain image is also used.
Next, the search processor 9 identifies one unprocessed DB brain image in the brain image DB 5 (step S65). Then, the search processor 9 calculates the degree of the similarity between the DB brain image identified at the step S65 and the query brain image, and stores the degree of the similarity into the storage device such as the main memory (step S67). At the step S67, it is possible that the difference between the density of a pixel of the query brain image and the density of a corresponding pixel of the DB brain image is calculated, and the total sum of the absolute values of the differences is calculated as the degree of the similarity. Incidentally, the pixel density difference between only portions in both of the query brain image and the DB brain image may be calculated.
In addition, it is possible that the image feature quantity of the color, the shape or the like is extracted from the query brain image to generate a feature quantity vector from the image feature quantity, the image feature quantity of the color, the shape or the like is extracted from the DB brain image to generate a feature quantity vector, and the distance between these feature quantity vectors is calculated as the degree of the similarity.
Then, the search processor 9 judges whether or not all of the DB brain image in the brain image DB 5 have been processed (step S69). When there is an unprocessed DB brain image, the processing returns to the step S65. On the other hand, when all of the DB brain image in the brain image DB 5 have been processed, the search processor 9 sorts the drawings corresponding to the DB brain images by using the data structure shown in
The input and output unit 13 presents the drawing corresponding to the DB brain image having the i-th degree of the similarity for the searcher, based on the sorting result stored in the search result storage 11 (step S73). For example, the screen including a “skip” button, a “select” button, and a “search again” button is displayed. Then, when an appropriate drawing is not displayed, the searcher clicks the “skip” button. When the “skip” is clicked, the input and output unit 13 accepts the skip instruction (step S75: Yes route), and judges whether or not “i” is the maximum value (step S81). When “i” is the maximum value, the processing shifts to a step S79. On the other hand, when “i” is not the maximum value, “i” is incremented (step S83), and the processing returns to the step S73.
On the other hand, when the “skip” button is not clicked (step S75: No route), but the “search again” button is clicked, the brain image input unit 7 accepts the re-search instruction (step S77: Yes route), and obtains a query brain image when the searcher imagines the drawing to be retrieved again after he or she watched the presented content, and stores the query brain image into the brain image storage 8 (step S79). Then, the input and output unit 13 initializes the counter i to “1” (step S80), and the processing returns to the step S65.
In addition, when the drawing the searcher visually checked or imagined is output, the searcher clicks the “select” button. The input and output unit 13 accepts the selection instruction of the drawing (step S77: No route), and outputs the drawing ID of the drawing to the brain image manager 3. Moreover, the input and output unit 13 instructs the search processor 9 to register the query brain image stored in the brain image storage 8 into the brain image DB 5, and the search processor 9 registers the query brain image into the brain image DB 5 in response to the instruction. In addition, the brain image manager 3 registers the drawing ID in association with the ID of the query brain image into the table as shown in
By carrying out the aforementioned processing, the query brain image can be refined to the image of the searcher step by step.
For example, by carrying out a following processing, the relation between the drawing and the brain image can be extracted to register it into the brain image DB 5 as new data.
The brain image manager 3 extracts an image feature quantity of the color, the shape or the like (e.g. an image feature quantity described in Takayuki Baba et al., “A shape-based part retrieval method for mechanical assembly drawings”, the technical study report of the Institute of Electronics, Information and Communication Engineers, PRMU2004-225, pp. 79-84 (2005)) from each drawing of the drawing DB 1, and arranges the DB brain images on the two-dimensional plane by using Self Organization Map (SOM) whose input data is the image feature quantity so that the DB brain images corresponding to the drawings having the similar image feature quantity gather nearby. Then, the brain image manager 3 presents the arrangement for the user (
The user selects the DB brain images within the frame 51, for example, from the presented content. Then, the brain image manager 3 accepts the selection of the DB brain images (step S93), identifies the drawings corresponding to the selected DB brain images by using the data structure as shown in
The user refers to the presented content as shown in
The brain image manager 3 accepts the input of the relation between the DB brain image and the drawing, and registers an ID of the DB brain image and an ID of the drawing into the data structure as shown in
Thus, the new relation can be registered into the brain image DB 5.
In order to reduce the measurement load of the brain image, when there are plural drawings having the extremely high degree of the similarity, only the representative drawing may be visually checked by the user to also assign the brain image to other similar drawings. For example, the brain image manager 3 calculates an image feature quantity for each drawing (e.g. the feature quantity in Takayuki Baba et al., “A shape-based part retrieval method for mechanical assembly drawings”, the technical study report of the Institute of Electronics, Information and Communication Engineers, PRMU2004-225, pp. 79-84 (2005)), and registers the image feature quantity into the drawing DB 1. Then, the brain image input unit 7 judges that the degree of the similarity is high when the distance between the calculated image feature quantities is less than a predesignated threshold, presents anyone of the drawings for the user to measure the brain image at that time, and outputs the obtained brain image to the brain image manager 3. The brain image manager 3 registers the data of the brain image measured this time into the brain image DB 5, and registers the drawing ID of the drawing presented for the user and the brain image ID of the brain image measured this time into the data structure shown in
The measurement workload of the brain image is reduced, and the brain image is associated with a lot of drawings.
In order to measure the query brain image, the fMRI measurement apparatus is required, for example. Therefore, there is a case where the query brain image cannot be easily obtained. For such a case, instead of the brain image input unit 7, the neural network may be used. The neural network is well known. Therefore, the details are not further explained.
In a case of this embodiment, as shown in
In addition, instead of the search processor 9 and the brain image DB 5, the neural network can be used. That is, as shown in
Thus, by using the neural network, the similar functions can be realized.
Conventionally, an input means should be investigated because there is a problem how embodies, as the query drawing, and efficiently inputs the query content intended by the searcher. By using the aforementioned search apparatus, when the searcher imagines, the brain image can be measured. Therefore, an effect in which the query drawing does not have to be prepared in advance can be obtained.
In addition, in the conventional similar shape search technique, the system developer has to determine the shape feature quantity used in the drawing search in advance, and the drawing search is carried out by using the similarity of the shape feature quantity. On the other hand, in this embodiment, the drawing search is carried out by using the similarity of the brain image measured when the searcher visually checks the drawing. Therefore, the secondary effect in which there is no need to determine the shape feature quantity in advance, and the similarity intended by the searcher can be easily added is realized.
Although the embodiments of this invention were explained, this invention is not limited to these embodiments. For example, the functional block diagram shown in
Incidentally, the search apparatus in this embodiment is a computer device as shown in
Although the present invention has been described with respect to a specific preferred embodiment thereof, various change and modifications may be suggested to one skilled in the art, and it is intended that the present invention encompass such changes and modifications as fall within the scope of the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2006-282519 | Oct 2006 | JP | national |
| Number | Name | Date | Kind |
|---|---|---|---|
| 6121969 | Jain et al. | Sep 2000 | A |
| 6798898 | Patton et al. | Sep 2004 | B1 |
| 6804683 | Matsuzaki et al. | Oct 2004 | B1 |
| 7236652 | Kasutani | Jun 2007 | B2 |
| Number | Date | Country | |
|---|---|---|---|
| 20080089553 A1 | Apr 2008 | US |
