Author: claimsbot

1-16. (canceled)
17. A semiconductor laser apparatus comprising:
a first semiconductor laser device having on a first substrate a first semiconductor layer that emits a light beam with a first wavelength; and
a second semiconductor laser device having on a second substrate a second semiconductor layer that emits a light beam with a second wavelength, wherein
said first and second wavelengths are different from each other, and materials of said first and second substrates are different from each other,
said second semiconductor laser device is laminated on said first semiconductor laser device so as not to overlap with a light-beam-emission point of said first semiconductor laser device in a vertical direction to one surface, on which said first semiconductor layer is formed, of said first substrate,
said first semiconductor laser device has a difference in level formed by an upper level surface and a lower level surface, the light-beam-emission point of said first semiconductor layer being arranged below said upper level surface, and
said second semiconductor laser device is laminated on the lower level surface of said first semiconductor laser device.
18. The semiconductor laser apparatus according to claim 17, wherein
said second semiconductor laser device is laminated on said first semiconductor laser device such that said second semiconductor layer side is positioned on said first semiconductor layer side.
19. The semiconductor laser apparatus according to claim 17, wherein
either of said first semiconductor layer and said second semiconductor layer is made of a nitride-based semiconductor.
20. The semiconductor laser apparatus according to claim 17, wherein
said first substrate is an optically transparent substrate.
21. The semiconductor laser apparatus according to claim 17, wherein
Said second semiconductor laser device is laminated on said first semiconductor laser device such that said first semiconductor layer is positioned on said second semiconductor laser device side.
22. The semiconductor laser apparatus according to claim 17, wherein
either of said first semiconductor layer and said second semiconductor layer includes a gallium arsenide-based semiconductor or a gallium indium phosphide-based semiconductor.
23. A semiconductor laser apparatus comprising:
a first semiconductor laser device having on a first substrate a first semiconductor layer that emits a light beam with a first wavelength; and
a second semiconductor laser device having on a second substrate a second semiconductor layer that emits a light beam with a second wavelength, wherein
said first and second wavelengths are different from each other, and materials of said first and second substrates are different from each other,
said second semiconductor laser device is laminated on said first semiconductor laser device so as not to overlap with a light-beam-emission point of said first semiconductor laser device in a vertical direction to one surface, on which said first semiconductor layer is formed, of said first substrate,
a heat dissipator is arranged in contact with a region on said first semiconductor laser device which overlaps with the light-beam-emission point of said first semiconductor layer in the vertical direction to said one surface of said first substrate and a surface of said second semiconductor laser device on the opposite side of said first semiconductor laser device.
24. The semiconductor laser apparatus according to claim 23, wherein
said second semiconductor laser device is laminated on said first semiconductor laser device, so that one surface of said first semiconductor laser device and one surface of said second semiconductor laser device form a difference in level, and
said heat dissipator is provided with a difference in level formed by a first surface in contact with the one surface of said first semiconductor laser device and a second surface in contact with the one surface of said second semiconductor laser device.
25. A semiconductor laser apparatus comprising:
a first semiconductor laser device having on a first substrate a first semiconductor layer that emits a light beam with a first wavelength;
a second semiconductor laser device having on a second substrate a second semiconductor layer that emits a light beam with a second wavelength; and
a third semiconductor laser device having a third semiconductor layer on a third substrate that emits a light beam with a third wavelength, wherein
said first and second wavelengths are different from each other, and materials of said first and second substrates are different from each other,
said second semiconductor laser device is laminated on said first semiconductor laser device so as not to overlap with a light-beam-emission point of said first semiconductor laser device in a vertical direction to one surface, on which said first semiconductor layer is formed, of said first substrate,
said third semiconductor laser device is laminated on said first semiconductor laser device except a region that overlaps with the light-beam-emission point of said first semiconductor laser device in the vertical direction to said one surface of said first substrate.
26. The semiconductor laser apparatus according to claim 25, wherein
said second and said third semiconductor laser devices are laminated on said first semiconductor laser device such that said first semiconductor layer is positioned on said second and said third semiconductor laser devices sides.
27. The semiconductor laser apparatus according to claim 25, wherein
said second semiconductor laser device is laminated on said first semiconductor laser device such that said second semiconductor layer side is positioned on said first semiconductor layer side.
28. The semiconductor laser apparatus according to claim 25, wherein
said third semiconductor laser device is laminated on said first semiconductor laser device such that said third semiconductor layer side is positioned on said first semiconductor layer side.
29. The semiconductor laser apparatus according to claim 25, wherein
said first, said second, and said third wavelengths are different from one another, and said first, said second, and said third semiconductor layers include any of a nitride-based semiconductor, a gallium arsenide-based semiconductor or a gallium indium phosphide-based semiconductor.
30. The semiconductor laser apparatus according to claim 25, wherein
A heat dissipator is arranged in contact with a region on said first semiconductor laser device with overlaps with the light-beam-emission point of said first semiconductor layer, a surface of said second semiconductor laser device on the opposite side of said first semiconductor laser device, and a surface of said third semiconductor laser device on the opposite side of said first semiconductor laser device.
31. A method of fabricating a semiconductor laser apparatus comprising the steps of:
forming on a first substrate a first semiconductor layer such that said first semiconductor layer has a plurality of first light-beam-emission points that emit light beams with a first wavelength;
forming a second semiconductor layer on a second substrate made of a different material from that of said first substrate such that said second semiconductor layer has a plurality of second light-beam-emission points that emit light beams with a second wavelength different from said first wavelength;
bonding said first substrate and said second substrate such that said second semiconductor layer is laminated on said first semiconductor layer;
etching said second substrate and said second semiconductor layer such that regions of said first semiconductor layer above said plurality of first light-beam-emission points become exposed; and
dividing a layered structure of said first semiconductor layer, said second substrate, and said second semiconductor layer into a plurality of semiconductor laser apparatuses.

The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.

1. A magnetic toner which internally contains magnetic particles and develops an electrostatic latent image, wherein:
said magnetic particle is of average particle size 0.01 \u03bcm through 0.50 \u03bcm;
said magnetic particle is of octahedron shape that is a convex polyhedron surrounded by eight triangles as a basis, each of vertexes and edges of the octahedron being in a curved surface shape and having a portion that can be taken as a straight line on the outer periphery of a projected image of said octahedron;
a sphericity of said magnetic toner is 0.94 through 0.98; and
a number content of said toner of particle size 0.6 \u03bcm through 2.0 \u03bcm is smaller than 10%.
2. The magnetic toner according to claim 1, wherein a content of said magnetic particle in said magnetic toner is 35 wt. % through 60 wt. %.
3. The magnetic toner according to claim 1, wherein a first shape of said magnetic particle is of octahedron shape that is a convex polyhedron surrounded by eight triangles as a basis, each of vertexes and edges of the octahedron being in a curved surface shape and having a portion that can be taken as a straight line on the outer periphery of a projected image of said octahedron, and a second shape of said magnetic particle is of not-tapered hexahedron or not-tapered octahedron, or tapered hexhedron or tapered octahedron wherein edges are tapered,
wherein: said magnetic particles of said first and second shape are mixed in said magnetic toner; and
a magnetization of said magnetic toner is equal to or greater than 2.0 Am2kg and smaller than 9.0 Am2.kg at a magnetic field 79.6 kAm.
4. An electro-photographic image forming method, using a magnetic toner which internally contains magnetic particles and develops an electrostatic latent image, wherein said magnetic particle is of average particle size 0.01 \u03bcm through 0.50 \u03bcm; said magnetic particle is of octahedron shape that is a convex polyhedron surrounded by eight triangles as a basis, each of vertexes and edges of the octahedron being in a curved surface shape and having a portion that can be taken as a straight line on the outer periphery of a projected image of said octahedron; a sphericity of said magnetic toner is 0.94 through 0.98; and a number content of said toner of particle size 0.6 \u03bcm through 2.0 \u03bcm is smaller than 10%, which comprises the steps of:
retaining magnetic toners on a rotating toner retaining member wherein magnets are fixed;
facing said developer retaining member, at a gap, against a latent image retaining member which is an a-silicon photoreceptor; and
flying said magnetic toners toward said latent image retaining member, thereby developing said latent image.

What is claimed is:

1. A method of forming trench isolation in a semiconductor device comprising:
forming a trench mask on a semiconductor substrate to define a trench forming region;
etching said semiconductor substrate using said trench mask and forming a trench therein;
forming a thermal oxide layer on a bottom and sidewalls of said trench, said thermal oxide layer removing substrate damage caused by said etching of said semiconductor substrate;
forming an oxidation barrier layer on said thermal oxide layer, said oxidation barrier layer preventing said bottom and sidewalls of said trench from being oxidized;
forming a protection layer on said oxidation barrier layer;
plasma processing said bottom and sidewalls of said trench; and
filling up said trench with a trench fill material uniformly with respect to said bottom and sidewalls of said trench.
2. The method according to claim 1, wherein said oxidation barrier layer comprises an LPCVD silicon nitride layer.
3. The method according to claim 2, wherein said silicon nitride layer comprises a silicon-rich nitride layer having greater silicon content as respectively compared to other component contents.
4. The method according to claim 1, wherein said protection layer comprises an oxide layer formed by deposition techniques at high temperature.
5. The method according to claim 4, wherein said protection layer is one selected from a group consisting of an HTO oxide layer and an LP-TBOS oxide layer.
6. The method according to claim 1, wherein said protection layer has a thickness of about 10 to 1,000 .
7. The method according to claim 1, further comprising, after said filling up said trench with said trench fill material, annealing so as to densify said trench fill material.
8. The method according to claim 7, wherein said annealing is carried out at a temperature of at least 900 C. in N2 atmosphere.
9. The method according to claim 7, wherein said step of annealing is carried out at a temperature of at least 700 C. in a wet atmosphere.
10. A trench isolation structure comprising:
a trench formed in a semiconductor substrate, said trench having sidewalls and a bottom;
a first and a second oxide layer formed on said bottom and sidewalls of said trench and a nitride layer disposed therebetween; and
a third oxide layer formed on said second oxide layer as to fill up said trench.
11. The trench isolation structure according to claim 10, wherein said nitride layer is an oxidation barrier layer, preventing said bottom and sidewalls of said trench from being oxidized, and further is a stress relief layer, relieving stress applied to said bottom and sidewalls of said trench.
12. The trench isolation structure according to claim 10, wherein said silicon nitride layer comprises a silicon-rich nitride layer having greater silicon content as respectively compared to other component contents.
13. The trench isolation structure according to claim 10, wherein said second oxide layer is one selected from a group consisting of an HTO oxide layer and an LP-CVD oxide layer.
14. The trench isolation structure according to claim 10, wherein said second oxide layer has a thickness of about 10 to 1,000 .

The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.

1. An image processing device comprising:
a studying image acquisition unit configured to acquire a studying image group including an image pair of a first-quality image and a second-quality image with different image qualities;
a tentative eigenprojection matrix generation unit configured to generate a provisional tentative eigenprojection matrix from the acquired studying image group by a projection computation;
a tentative projection core tensor creation unit configured to create a tentative projection core tensor defining a correspondence between the first-quality image and an intermediate eigenspace and a correspondence between the second-quality image and the intermediate eigenspace from the acquired studying image group and the generated tentative eigenprojection matrix;
a first tentative sub-core tensor creation unit configured to create a first tentative sub-core tensor corresponding to a condition specified by a first setting from the created tentative projection core tensor;
a first tentative subtensor projection unit configured to project at least one of the first-quality image and the second-quality image in the studying image group by a first projection computation using the tentative eigenprojection matrix and the first tentative sub-core tensor to calculate a coefficient vector in the intermediate eigenspace;
a representative vector calculation unit configured to obtain a representative intermediate eigenspace coefficient vector, which represents the coefficient vector group, from the coefficient vector group calculated by the first tentative subtensor projection unit in accordance with the number of studying representatives;
a representative studying image group determination unit configured to determine a representative studying image group from the representative intermediate eigenspace coefficient vector obtained by the representative vector calculation unit;
an eigenprojection matrix generation unit configured to generate an eigenprojection matrix from the determined representative studying image group by the projection computation; and
a projection core tensor creation unit configured to create a projection core tensor defining a correspondence between the first-quality image and the intermediate eigenspace and a correspondence between the second-quality image and the intermediate eigenspace from the determined representative studying image group and the eigenprojection matrix generated by the eigenprojection matrix generation unit.
2. The image processing device according to claim 1, comprising
a number of studying representatives acquisition unit configured to acquire information on the number of studying representatives from the outside.
3. The image processing device according to claim 1, comprising
a storage unit configured to store the eigenprojection matrix generated by the eigenprojection matrix generation unit and the projection core tensor created by the projection core tensor creation unit.
4. The image processing device according to claim 3, comprising:
a first sub-core tensor creation unit configured to create a first sub-core tensor corresponding to a condition specified by the first setting from the projection core tensor created by the projection core tensor creation unit; and
a second sub-core tensor creation unit configured to create a second sub-core tensor corresponding to a condition specified by a second setting from the projection core tensor, wherein
the first sub-core tensor created by the first sub-core tensor creation unit and the second sub-core tensor created by the second sub-core tensor creation unit are stored and preserved in the storage unit in place of the projection core tensor or along with the projection core tensor.
5. The image processing device according to claim 1 comprising:
a first sub-core tensor creation unit configured to create a first sub-core tensor corresponding to a condition specified by the first setting from the created projection core tensor;
a second sub-core tensor creation unit configured to create a second sub-core tensor corresponding to a condition specified by a second setting from the created projection core tensor;
a first subtensor projection unit configured to project an input image to be processed by a first projection computation using the eigenprojection matrix and the first sub-core tensor to calculate a coefficient vector in the intermediate eigenspace; and
a second subtensor projection unit configured to project the calculated coefficient vector by a second projection computation using the second sub-core tensor and the eigenprojection matrix to generate a change image with an image quality different from the input image.
6. The image processing device according to claim 5, comprising:
a first feature area specifying unit configured to specify a first feature area from an inputted image;
a compression processing unit configured to compress an image part of the first feature area in the inputted image at a first compressive strength and to compress an image part of other than the first feature area at a second compressive strength which is a compressive strength higher than the first compressive strength; and
a unit configured to project at least the first feature area by the first subtensor projection unit and the second subtensor projection unit to change the image quality.
7. An image processing method comprising:
a studying image acquisition step of acquiring a studying image group including an image pair of a first-quality image and a second-quality image with different image qualities;
a tentative eigenprojection matrix generation step of generating a provisional tentative eigenprojection matrix from the acquired studying image group by a projection computation;
a tentative projection core tensor creation step of creating a tentative projection core tensor defining a correspondence between the first-quality image and an intermediate eigenspace and a correspondence between the second-quality image and the intermediate eigenspace from the acquired studying image group and the generated tentative eigenprojection matrix;
a first tentative sub-core tensor creation step of creating a first tentative sub-core tensor corresponding to a condition specified by a first setting from the created tentative projection core tensor;
a first tentative subtensor projection step of projecting at least one of the first-quality image and the second-quality image in the studying image group by a first projection computation using the tentative eigenprojection matrix and the first tentative sub-core tensor to calculate a coefficient vector in the intermediate eigenspace;
a representative vector calculation step of obtaining a representative intermediate eigenspace coefficient vector, which represents the coefficient vector group, from the coefficient vector group calculated by the first tentative subtensor projection in accordance with the number of studying representatives;
a representative studying image group determination step of determining a representative studying image group from the representative intermediate eigenspace coefficient vector obtained in the representative vector calculation step;
an eigenprojection matrix generation step of generating an eigenprojection matrix from the determined representative studying image group by the projection computation; and
a projection core tensor creation step of creating a projection core tensor defining a correspondence between the first-quality image and the intermediate eigenspace and a correspondence between the second-quality image and the intermediate eigenspace from the determined representative studying image group and the eigenprojection matrix generated in the eigenprojection matrix generation step.
8. The image processing method according to claim 7, comprising
a number of studying representatives acquisition step of acquiring information on the number of studying representatives from the outside.
9. The image processing method according to claim 7, comprising
a storage step of storing, in storage unit, the eigenprojection matrix generated in the eigenprojection matrix generation step and the projection core tensor created in the projection core tensor creation step.
10. The image processing method according to claim 9, comprising:
a first sub-core tensor creation step of creating a first sub-core tensor corresponding to a condition specified by the first setting from the projection core tensor created in the projection core tensor creation step; and
a second sub-core tensor creation step of creating a second sub-core tensor corresponding to a condition specified by a second setting from the projection core tensor, wherein
the first sub-core tensor created in the first sub-core tensor creation step and the second sub-core tensor created in the second sub-core tensor creation step are stored and preserved in the storage unit in place of the projection core tensor or along with the projection core tensor.
11. The image processing method according to claim 7 comprising:
a first sub-core tensor creation step of creating a first sub-core tensor corresponding to a condition specified by the first setting from the created projection core tensor;
a second sub-core tensor creation step of creating a second sub-core tensor corresponding to a condition specified by a second setting from the created projection core tensor;
a first subtensor projection step of projecting an input image to be processed by a first projection computation using the eigenprojection matrix and the first sub-core tensor to calculate a coefficient vector in the intermediate eigenspace; and
a second subtensor projection step of projecting the calculated coefficient vector by a second projection computation using the second sub-core tensor and the eigenprojection matrix to generate a change image with an image quality different from the input image.
12. A non-transitory computer-readable recording medium including a program stored thereon, such that when the program is read and executed by a computer, the computer is caused to function as:
a studying image acquisition unit configured to acquire a studying image group including an image pair of a first-quality image and a second-quality image with different image qualities;
a tentative eigenprojection matrix generation unit configured to generate a provisional tentative eigenprojection matrix from the acquired studying image group by a projection computation;
a tentative projection core tensor creation unit configured to create a tentative projection core tensor defining a correspondence between the first-quality image and an intermediate eigenspace and a correspondence between the second-quality image and the intermediate eigenspace from the acquired studying image group and the generated tentative eigenprojection matrix;
a first tentative sub-core tensor creation unit configured to create a first tentative sub-core tensor corresponding to a condition specified by a first setting from the created tentative projection core tensor;
a first tentative subtensor projection unit configured to project at least one of the first-quality image and the second-quality image in the studying image group by a first projection computation using the tentative eigenprojection matrix and the first tentative sub-core tensor to calculate a coefficient vector in the intermediate eigenspace;
a representative vector calculation unit configured to obtain a representative intermediate eigenspace coefficient vector, which represents the coefficient vector group, from the coefficient vector group calculated by the first tentative subtensor projection unit in accordance with the number of studying representatives;
a representative studying image group determination unit configured to determine a representative studying image group from the representative intermediate eigenspace coefficient vector obtained by the representative vector calculation unit;
an eigenprojection matrix generation unit configured to generate an eigenprojection matrix from the determined representative studying image group by the projection computation; and
a projection core tensor creation unit configured to create a projection core tensor defining a correspondence between the first-quality image and the intermediate eigenspace and a correspondence between the second-quality image and the intermediate eigenspace from the determined representative studying image group and the eigenprojection matrix generated by the eigenprojection matrix generation unit.
13. The medium according to claim 12, further causing the computer to function as
a number of studying representatives acquisition unit configured to acquire information on the number of studying representatives from the outside.
14. The medium according to claim 12, further causing the computer to function as
a storage unit configured to store the eigenprojection matrix generated by the eigenprojection matrix generation unit and the projection core tensor created by the projection core tensor creation unit.
15. The medium according to claim 12, further causing the computer to function as:
a first sub-core tensor creation unit configured to create a first sub-core tensor corresponding to a condition specified by the first setting from the projection core tensor created by the projection core tensor creation unit; and
a second sub-core tensor creation unit configured to create a second sub-core tensor corresponding to a condition specified by a second setting from the projection core tensor, wherein
the first sub-core tensor created by the first sub-core tensor creation unit and the second sub-core tensor created by the second sub-core tensor creation unit are stored and preserved in the storage unit in place of the projection core tensor or along with the projection core tensor.
16. The medium according to claim 12, further causing the computer to function as:
a first sub-core tensor creation unit configured to create a first sub-core tensor corresponding to a condition specified by the first setting from the created projection core tensor;
a second sub-core tensor creation unit configured to create a second sub-core tensor corresponding to a condition specified by a second setting from the created projection core tensor;
a first subtensor projection unit configured to project an input image to be processed by a first projection computation using the eigenprojection matrix and the first sub-core tensor to calculate a coefficient vector in the intermediate eigenspace; and
a second subtensor projection unit configured to project the calculated coefficient vector by a second projection computation using the second sub-core tensor and the eigenprojection matrix to generate a change image with an image quality different from the input image.
17. A data processing device comprising:
a tentative eigenprojection matrix generation unit configured to generate, by a projection computation, a provisional tentative eigenprojection matrix from a studying data group including a pair of first-condition data and second-condition data with different conditions;
a tentative projection core tensor creation unit configured to create a tentative projection core tensor defining a correspondence between the first-condition data and an intermediate eigenspace and a correspondence between the second-condition data and the intermediate eigenspace from the studying data group and the generated tentative eigenprojection matrix;
a first tentative sub-core tensor creation unit configured to create a first tentative sub-core tensor corresponding to a condition specified by a first setting from the created tentative projection core tensor;
a first tentative subtensor projection unit configured to project an image of at least one of the first-condition data and the second-condition data in the studying data group by a first projection computation using the tentative eigenprojection matrix and the first tentative sub-core tensor to calculate a coefficient vector in the intermediate eigenspace;
a representative vector calculation unit configured to obtain a representative intermediate eigenspace coefficient vector, which represents the coefficient vector group, from the coefficient vector group calculated by the first tentative subtensor projection unit in accordance with the number of studying representatives;
a representative studying data group determination unit configured to determine a representative studying data group from the representative intermediate eigenspace coefficient vector obtained by the representative vector calculation unit;
a eigenprojection matrix generation unit configured to generate an eigenprojection matrix from the determined representative studying data group by the projection computation;
a projection core tensor creation unit configured to create a projection core tensor defining a correspondence between the first-condition data and the intermediate eigenspace and a correspondence between the second-condition data and the intermediate eigenspace from the determined representative studying data group and the eigenprojection matrix generated by the eigenprojection matrix generation unit; and
a storage unit configured to store the eigenprojection matrix generated by the eigenprojection matrix generation unit and the projection core tensor created by the projection core tensor creation unit.
18. A data processing method comprising:
a tentative eigenprojection matrix generation step of generating, by a projection computation, a provisional tentative eigenprojection matrix from a studying data group including a pair of first-condition data and second-condition data with different conditions;
a tentative projection core tensor creation step of creating a tentative projection core tensor defining a correspondence between the first-condition data and an intermediate eigenspace and a correspondence between the second-condition data and the intermediate eigenspace from the studying data group and the generated tentative eigenprojection matrix;
a first tentative sub-core tensor creation step of creating a first tentative sub-core tensor corresponding to a condition specified by a first setting from the created tentative projection core tensor;
a first tentative subtensor projection step of projecting an image of at least one of the first-condition data and the second-condition data in the studying data group by a first projection computation using the tentative eigenprojection matrix and the first tentative sub-core tensor to calculate a coefficient vector in the intermediate eigenspace;
a representative vector calculation step of obtaining a representative intermediate eigenspace coefficient vector, which represents the coefficient vector group, from the coefficient vector group calculated in the first tentative subtensor projection step in accordance with the number of studying representatives;
a representative studying data group determination step of determining a representative studying data group from the representative intermediate eigenspace coefficient vector obtained in the representative vector calculation step;
an eigenprojection matrix generation step of generating an eigenprojection matrix from the determined representative studying data group by the projection computation;
a projection core tensor creation step of creating a projection core tensor defining a correspondence between the first-condition data and the intermediate eigenspace and a correspondence between the second-condition data and the intermediate eigenspace from the determined representative studying data group and the eigenprojection matrix generated in the eigenprojection matrix generation step; and
a storage step of storing, in storage unit, the eigenprojection matrix generated in the eigenprojection matrix generation step and the projection core tensor created in the projection core tensor creation step.
19. A non-transitory computer-readable recording medium including a program stored thereon, such that when the program is read and executed by a computer, the computer is caused to function as:
a tentative eigenprojection matrix generation unit configured to generate, by a projection computation, a provisional tentative eigenprojection matrix from a studying data group including a pair of first-condition data and second-condition data with different conditions;
a tentative projection core tensor creation unit configured to create a tentative projection core tensor defining a correspondence between the first-condition data and an intermediate eigenspace and a correspondence between the second-condition data and the intermediate eigenspace from the studying data group and the generated tentative eigenprojection matrix;
a first tentative sub-core tensor creation unit configured to create a first tentative sub-core tensor corresponding to a condition specified by a first setting from the created tentative projection core tensor;
a first tentative subtensor projection unit configured to project an image of at least one of the first-condition data and the second-condition data in the studying data group by a first projection computation using the tentative eigenprojection matrix and the first tentative sub-core tensor to calculate a coefficient vector in the intermediate eigenspace;
a representative vector calculation unit configured to obtain a representative intermediate eigenspace coefficient vector, which represents the coefficient vector group, from the coefficient vector group calculated by the first tentative subtensor projection unit in accordance with the number of studying representatives;
a representative studying data group determination unit configured to determine a representative studying data group from the representative intermediate eigenspace coefficient vector obtained by the representative vector calculation unit;
an eigenprojection matrix generation unit configured to generate an eigenprojection matrix from the determined representative studying data group by the projection computation;
a projection core tensor creation unit configured to create a projection core tensor defining a correspondence between the first-condition data and the intermediate eigenspace and a correspondence between the second-condition data and the intermediate eigenspace from the determined representative studying data group and the eigenprojection matrix generated by the eigenprojection matrix generation unit; and
a storage unit configured to store the eigenprojection matrix generated by the eigenprojection matrix generation unit and the projection core tensor created by the projection core tensor creation unit.