What is claimed is:
1. An improved vacuum heat treating furnace having a water-cooled cylindrical vacuum-tight vessel with loading doors at either end for receiving a workload, pump means for evacuating the vessel, a cylindrical radiant heat-reflecting assembly concentrically offset from the interior of the vessel to form an internal hot zone and an annular channel with said vessel, a planar radiant heat-reflecting assembly offset from the interior of each loading door to form an end channel therewith, and inlet and outlet ports communicating with the chamber and the channel, the improvement comprising, in combination:
matching first pairs of elongate radiant energy heating units formed to be uniformly spaced in coaxial alignment around and offset from the interior of the heat-reflecting assembly, said units in each of said first pairs being located at mutually opposed sides of the heat-reflecting assembly and include two parallel spaced resistance elements electrically connected in series at adjacent ends thereof by a jumper plate; and
matching second pairs of elongate radiant energy heating units formed to be uniformly spaced in transverse alignment and offset from the interior of each of the planar assemblies, said units in each of said second pairs being located at mutually opposed sides of the planar assemblies and include parallel spaced resistance elements electrically connected in series; and
each of said elements including flat elongate sections with adjacent sections disposed from each other at an included angle for radiating energy inwardly and outwardly in diverse directions, the amount of said offset from the assemblies and said angle being selected to effect optimum direct and reflected energy into the chamber.
2. The improvement of claim 1 wherein said elements each comprise a middle section and opposed side sections with said included angles facing the adjacent one of said assemblies.
3. The improvement of claim 2 wherein said included angles are obtuse.
4. The improvement of claim 3 wherein the widths of said middle and side sections are equal.
5. The improvement of claim 4 wherein said elements are offset from said assemblies approximately two and one half times the width of one of said sections.
6. The improvement of claim 1 wherein said elements comprise two sections with said included angle facing the adjacent one of said assemblies.
7. The improvement of claim 6 wherein said included angle is about 90 degrees.
8. The improvement of claim 7 wherein the width of said sections are equal.
9. The improvement of claim 8 wherein said elements are offset from said assemblies approximately one and one half times the width of one of said sections.
10. The improvement of claim 1 further comprising:
means for introducing an inert cooling fluid into the furnace;
heat exchanger means operatively connected between the inlet and outlet ports;
blower means operatively connected for circulating the fluid through the furnace and said heat exchanger means;
first means responsive to the temperature of the fluid at said outlet for modulating the fluid flow only through the channels; and
second means responsive to the temperature of the fluid at said outlet for modulating the fluid flow only through the hot zones.
11. The improvement of claim 10 wherein:
said first means decreases the flow through the channels with decreasing temperature, and;
said second means increases the flow through the chamber with decreasing temperature.
12. Apparatus for an electric resistance heat treating furnace including a cylindrical heat shield assembly and a planar heat shield assembly at each end forming an interior chamber, comprising:
first pairs of elongated radiant energy heating units formed to be uniformly spaced in coaxial alignment around and offset from the interior of said cylindrical heat shield assembly, said units in each of said first pairs being located at mutually opposed sides of said cylindrical heat shield assembly and include parallel spaced resistance elements electrically connected in series at adjacent ends thereof; and
second pairs of elongated radiant energy heating units formed to be uniformly spaced in transverse alignment and offset from the interior of each of said planar heat shield assemblies, said units of each pair being located at mutually opposed sides of said planar assemblies and include parallel spaced resistance elements electrically connected in series at adjacent ends thereof;
each of said elements including relatively thin flat elongate sections angularly disposed from each other for radiating energy in diverse directions into the chamber and toward said assemblies.
13. The apparatus of claim 12 wherein adjacent ones of said sections each form an included angle facing the adjacent one of said assemblies.
14. The apparatus of claim 13 wherein the amount of said offset and said angles are selected to effect optimum direct and reflected energy into the chamber.
15. Apparatus for cooling an electric resistance heat-treating furnace comprising:
a cylindrical vacuum-tight vessel with loading doors at either end for receiving a workload, said vessel including inlet and outlet ports;
a cylindrical heat-reflecting assembly concentrically offset from the interior of said vessel defining an interior hot zone and an annular channel between said cylindrical assembly and said vessel;
a planar heat-reflecting assembly offset from the interior of each of said loading doors defining end channels at either end between said planar assembly and said doors;
first radiant heating units circumferentially spaced around and offset from the interior of said cylindrical assembly;
second radiant heating units transversely spaced and offset from the interior of said planar assemblies;
a heat exchanger operatively connected between an inlet and an outlet port of said vessel for cooling a fluid from said outlet port;
blower means operatively connected between said heat exchanger and said inlet port for circulating the fluid through said chamber and said channel;
first means responsive to the temperature of the fluid at said outlet port for modulating the fluid flow only through said channels; and
second means responsive to the temperature of the fluid at said outlet for modulating the fluid flow only through said chamber.
16. The improvement of claim 15 wherein:
said first means decreases the flow through said channels with decreasing temperature, and;
said second means increases the flow through said hot zone with decreasing temperature
17. An improved radiant heating unit for an electric heat treating furnace having radiant heat-reflecting assembly, the improvement comprising:
a plurality of elongate parallel-spaced resistance elements mounted on the assembly, each element including relatively thin flat radiant heating sections angularly disposed from each other for radiating energy in diverse directions; and
a jumper plate connected at mutually adjacent one ends of said sections connecting said sections in electrical series.
18. The improvement of claim 17 wherein each of said sections define a middle section and opposed side sections each formed with an included angle facing an interior surface of said assembly.
19. The improvement of claim 18 wherein said elements are offset from said assembly to effect optimum direct and reflected energy into the furnace.
20. The improvement of claim 17 wherein each of said sections define two sections with an included angle facing an interior surface of the assembly.
21. The improvement of claim 20 wherein said included angle and said offset from said assembly effect optimum direct and reflected energy into the furnace.
22. A method for cooling an electric heat treating furnace to a desired end temperature following a heat treating phase, the furnace including radiant heat-reflecting assemblies offset from the interior surface forming a hot zone within the assemblies and an annular channel in the space between the assemblies and the interior surface, fluid inlet and outlet ports communicating with the hot zone and channel, and a heat exchanger operatively connected between said inlet and outlet ports, comprising the sequential steps of:
backfilling the furnace with a cooling fluid;
recirculating the cooling fluid through the hot zone, channels and heat exchanger;
decreasing recirculation of the cooling fluid only through the channel as the temperature of the cooling fluid at the outlet port decreases; and
increasing recirculation of the cooling fluid only through the hot zone as the temperature of the cooling fluid at the outlet port decreases.
23. Apparatus for cooling an electric heat treating furnace to a desired end temperature following a heat treating phase, the furnace including a radiant heat-reflecting assembly offset from the interior surface of the furnace forming a hot zone within the assembly and a hot zone in a space between the assemblies and the interior surface, fluid inlet and outlet ports communicating with the hot zone and channel, and a heat exchanger operatively connected between the inlet and outlet ports, comprising:
means for introducing a cooling fluid into the furnace through the inlet port;
first regulator means responsive to the temperature of the fluid at the outlet port for decreasing the flow through only the channels with decreasing temperature; and
second regulator means responsive to the temperature of the fluid at the outlet port for increasing the flow through only the hot zone with decreasing temperature.
24. An improved electric resistance radiant heating unit, comprising:
at least two elongate flat sections angularly disposed from each other.
25. The improvement of claim 24 wherein said sections comprise a middle section and opposed side sections each forming with said middle section an included angle.
26. The improvement of claim 24 wherein sections comprise two sections of substantially equal width forming an included angle.
The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.
What is claimed is:
1. An image retrieval apparatus which retrieves an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
first image feature vector conversion means for receiving similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation, and converting the image feature vector of the query image into a first image feature vector according to the similarity computation image feature vector configuration information; and
second image feature vector conversion means for converting the image feature vector of the images to be retrieved into a second image feature vector according to the similarity computation image feature vector configuration information,
wherein similarity between the first image feature vector and the second image feature vector is computed according to the similarity computation image feature vector configuration information.
2. An image retrieval apparatus which retrieves an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
first configuration information prospect selection means for referring to query image feature vector configuration information describing a configuration of an image feature vector of the query image, and outputting a configuration information list of possible feature vector configurations of the image feature vector of the query image;
second configuration information prospect selection means for referring to images-to-be-retrieved feature vector configuration information describing the configuration of an image feature vector of the image to be retrieved, and outputting a configuration information list of possible feature vector configurations of the image feature vectors of the images to be retrieved;
configuration information discrimination means for receiving similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation, and determining whether or not the similarity computation image feature vector configuration information is contained in both configuration information list obtained from said first configuration information prospect selection means and a configuration information list obtained from said second configuration information prospect selection means;
first image feature vector conversion means for converting the image feature vector of the query image into a first image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said configuration information discrimination means determines that the information is contained in both lists; and
second image feature vector conversion means for converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said configuration information discrimination means determines that the information is contained in both lists,
wherein similarity between the first image feature vector and the second image feature vector is computed according to the similarity computation image feature vector configuration information.
3. An image retrieval apparatus which retrieves an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
determination means for determining, according to similarity computation image feature vector configuration information given as input or similarity computation image feature vector configuration information generated by configuration information reselection means, whether or not both of image feature vector of the query image and image feature vectors of the images to be retrieved can be converted;
said configuration information reselection means for selecting similarity computation image feature vector configuration information again when said determination means determines that the image feature vectors cannot be converted;
first image feature vector conversion means for converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information when said determination means determines that the image feature vectors can be converted; and
second image feature vector conversion means for converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said determination means determines that the image feature vectors can be converted,
wherein similarity between the first image feature vector and the second image feature vector is computed according to the similarity computation image feature vector configuration information.
4. An image retrieval apparatus which retrieves an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
configuration information generation means for generating similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
first image feature vector conversion means for converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information; and
second image feature vector conversion means for converting the image feature vectors of the images to be retrieved into a second image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information,
wherein similarity between the first image feature vector and the second image feature vector is computed according to the similarity computation image feature vector configuration information.
5. An image retrieval apparatus which retrieves an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
first configuration information prospect selection means for referring to query image feature vector configuration information describing a configuration of an image feature vector of the query image, and outputting a configuration information list of possible feature vector configurations of the image feature vector of the query image;
second configuration information prospect selection means for referring to images-to-be-retrieved feature vector configuration information describing the configuration of an image feature vector of the image to be retrieved, and outputting a configuration information list of possible feature vector configurations of the image feature vectors of the images to be retrieved;
configuration information determination means for referring to the configuration information list obtained by said first configuration information prospect selection means and the configuration information list obtained by said second configuration information prospect selection means, and determining similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
first image feature vector conversion means for converting the image feature vector of the query image into a first image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information; and
second image feature vector conversion means for converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information,
wherein similarity between the first image feature vector and the second image feature vector is computed according to the similarity computation image feature vector configuration information.
6. An image retrieval apparatus which retrieves an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
configuration information generation means for generating similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
determination means for determining, according to similarity computation image feature vector configuration information generated by said configuration information generation means or configuration information reselection means, whether or not both of image feature vector of the query image and image feature vectors of the images to be retrieved can be converted;
said configuration information reselection means for selecting similarity computation image feature vector configuration information again when said determination means determines that the image feature vectors cannot be converted;
first image feature vector conversion means for converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information when said determination means determines that the image feature vectors can be converted; and
second image feature vector conversion means for converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said determination means determines that the image feature vectors can be converted,
wherein similarity between the first image feature vector and the second image feature vector is computed according to the similarity computation image feature vector configuration information.
7. The image retrieval apparatus according to claim 4, wherein said configuration information generation means refers to query image feature vector configuration information describing a configuration of the image feature vector of the query image and images-to-be-retrieved feature vector configuration information describing a configuration of the image feature vectors of the images to be retrieved, and generates similarity computation image feature vector configuration information.
8. The image retrieval apparatus according to claim 1, wherein when similarity between the first image feature vector and the second image feature vector is computed, normalized similarity can be computed depending on contents of the similarity computation image feature vector configuration information.
9. The image retrieval apparatus according to claim 1, wherein the image feature vector can be generated by obtaining a reduced image from an image of which a feature vector is to be extracted, and performing discrete cosine transform and quantization on the reduced image.
10. The image retrieval apparatus according to claim 1, wherein the similarity computation image feature vector configuration information relates to a configuration of discrete cosine transform coefficients forming an image feature vector.
11. The image retrieval apparatus according to claim 1, wherein the image feature vector is one or more dominant colors of an image, and the similarity computation image feature vector configuration information relates to the dominant color(s).
12. The image retrieval apparatus according to claim 1, wherein the image feature vector indicates a direction of an edge of an image, and the similarity computation image feature vector configuration information relates to a type of the direction of the edge.
13. The image retrieval apparatus according to claim 1, wherein the image feature vector indicates a texture of an image, and the similarity computation image feature vector configuration information relates to a type of the texture.
14. The image retrieval apparatus according to claim 1, wherein the image feature vector indicates a shape of an image, and the similarity computation image feature vector configuration information relates to a type of the shape.
15. The image retrieval apparatus according to claim 1, wherein the image feature vector indicates a motion of an image, and the similarity computation image feature vector configuration information relates to a type of the motion.
16. The image retrieval apparatus according to claim 1, wherein the image feature vector is generated by extracting a feature vector of each rectangular area after dividing an image into rectangular areas, and the similarity computation image feature vector configuration information relates to the rectangular area.
17. An image retrieving method of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a first image feature vector conversion step of receiving similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation, and converting the image feature vector of the query image into a first image feature vector according to the similarity computation image feature vector configuration information;
a second image feature vector conversion step of converting the image feature vector of the images to be retrieved into a second image feature vector according to the similarity computation image feature vector configuration information; and
a step of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
18. An image retrieving method of retrieving an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
a first configuration information prospect selection step of referring to query image feature vector configuration information describing a configuration of an image feature vector of the query image, and outputting a configuration information list of possible feature vector configurations of the image feature vector of the query image;
a second configuration information prospect selection step of referring to images-to-be-retrieved feature vector configuration information describing the configuration of an image feature vector of the image to be retrieved, and outputting a configuration information list of possible feature vector configurations of the image feature vectors of the images to be retrieved;
a configuration information discrimination step of receiving similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation, and determining whether or not the similarity computation image feature vector configuration information is contained in both configuration information list obtained in said first configuration information prospect selection step and a configuration information list obtained in said second configuration information prospect selection step;
a first image feature vector conversion step of converting the image feature vector of the query image into a first image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said configuration information discrimination step determines that the information is contained in both lists;
a second image feature vector conversion step of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said configuration information discrimination step determines that the information is contained in both lists; and
a step of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
19. An image retrieving method of retrieving an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
a determination step of determining, according to similarity computation image feature vector configuration information given as input or similarity computation image feature vector configuration information generated in a configuration information reselection step, whether or not both of image feature vector of the query image and image feature vectors of the images to be retrieved can be converted;
said configuration information reselection step of selecting similarity computation image feature vector configuration information again when said determination step determines that the image feature vectors cannot be converted;
a first image feature vector conversion step of converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information when said determination step determines that the image feature vectors can be converted;
a second image feature vector conversion step of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said determination step determines that the image feature vectors can be converted; and
a step of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
20. An image retrieving method of retrieving an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
a configuration information generation step of generating similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
a first image feature vector conversion step of converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information;
a second image feature vector conversion step of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information; and
a step of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
21. An image retrieving method of retrieving an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
a first configuration information prospect selection step of referring to query image feature vector configuration information describing a configuration of an image feature vector of the query image, and outputting a configuration information list of possible feature vector configurations of the image feature vector of the query image;
a second configuration information prospect selection step of referring to images-to-be-retrieved feature vector configuration information describing the configuration of an image feature vector of the image to be retrieved, and outputting a configuration information list of possible feature vector configurations of the image feature vectors of the images to be retrieved;
a configuration information determination step of referring to the configuration information list obtained in said first configuration information prospect selection step and the configuration information list obtained in said second configuration information prospect selection step, and determining similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
a first image feature vector conversion step of converting the image feature vector of the query image into a first image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information;
a second image feature vector conversion step of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information; and
a step of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
22. An image retrieving method of retrieving an image similar to a predetermined query image from images to be retrieved by using an image feature vector, comprising:
a configuration information generation step of generating similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
a determination step of determining, according to similarity computation image feature vector configuration information generated in said configuration information generation step or configuration information reselection step, whether or not both of image feature vector of the query image and image feature vectors of the images to be retrieved can be converted;
said configuration information reselection step of selecting similarity computation image feature vector configuration information again when said determination step determines that the image feature vectors cannot be converted;
a first image feature vector conversion step of converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information when said determination step determines that the image feature vectors can be converted;
a second image feature vector conversion step of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said determination step determines that the image feature vectors can be converted; and
a step of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
23. The image retrieving method according to claim 20 or 22, wherein said configuration information generation step refers to query image feature vector configuration information describing a configuration of the image feature vector of the query image and images-to-be-retrieved feature vector configuration information describing a configuration of the image feature vectors of the images to be retrieved, and generates similarity computation image feature vector configuration information.
24. The image retrieving method according to claim 17, wherein when similarity between the first image feature vector and the second image feature vector is computed, normalized similarity can be computed depending on contents of the similarity computation image feature vector configuration information.
25. The image retrieving method according to claim 17, wherein the image feature vector can be generated by obtaining a reduced image from an image of which a feature vector is to be extracted, and performing discrete cosine transform and quantization on the reduced image.
26. The image retrieving method according to claim 17, wherein the similarity computation image feature vector configuration information relates to a configuration of discrete cosine transform coefficients forming an image feature vector.
27. The image retrieving method according to claim 17, wherein the image feature vector is one or more dominant colors of an image, and the similarity computation image feature vector configuration information relates to the dominant color(s).
28. The image retrieving method according to claim 17, wherein the image feature vector indicates a direction of an edge of an image, and the similarity computation image feature vector configuration information relates to a type of the direction of the edge.
29. The image retrieving method according to claim 17, wherein the image feature vector indicates a texture of an image, and the similarity computation image feature vector configuration information relates to a type of the texture.
30. The image retrieving method according to claim 17, wherein the image feature vector indicates a shape of an image, and the similarity computation image feature vector configuration information relates to a type of the shape.
31. The image retrieving method according to claim 17, wherein the image feature vector indicates a motion of an image, and the similarity computation image feature vector configuration information relates to a type of the motion.
32. The image retrieving method according to claim 17, wherein the image feature vector is generated by extracting a feature vector of each rectangular area after dividing an image into rectangular areas, and the similarity computation image feature vector configuration information relates to the rectangular area.
33. An image retrieving program for directing a computer to perform an image retrieving process of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a first image feature vector conversion process of receiving similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation, and converting the image feature vector of the query image into a first image feature vector according to the similarity computation image feature vector configuration information;
a second image feature vector conversion process of converting the image feature vector of the images to be retrieved into a second image feature vector according to the similarity computation image feature vector configuration information; and
a process of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
34. An image retrieving program for directing a computer to perform an image retrieving process of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a first configuration information prospect selection process of referring to query image feature vector configuration information describing a configuration of an image feature vector of the query image, and outputting a configuration information list of possible feature vector configurations of the image feature vector of the query image;
a second configuration information prospect selection process of referring to images-to-be-retrieved feature vector configuration information describing the configuration of an image feature vector of the image to be retrieved, and outputting a configuration information list of possible feature vector configurations of the image feature vectors of the images to be retrieved;
a configuration information discrimination process of receiving similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation, and determining whether or not the similarity computation image feature vector configuration information is contained in both configuration information list obtained in said first configuration information prospect selection process and a configuration information list obtained in said second configuration information prospect selection process;
a first image feature vector conversion process of converting the image feature vector of the query image into a first image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said configuration information discrimination process determines that the information is contained in both lists;
a second image feature vector conversion process of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said configuration information discrimination process determines that the information is contained in both lists; and
a process of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
35. An image retrieving program for directing a computer to perform an image retrieving process of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a determination process of determining, according to similarity computation image feature vector configuration information given as input or similarity computation image feature vector configuration information generated in a configuration information reselection process, whether or not both of image feature vector of the query image and image feature vectors of the images to be retrieved can be converted;
a configuration information reselection process of selecting similarity computation image feature vector configuration information again when said determination process determines that the image feature vectors cannot be converted;
a first image feature vector conversion process of converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information when said determination process determines that the image feature vectors can be converted;
a second image feature vector conversion process of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said determination process determines that the image feature vectors can be converted; and
a process of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
36. An image retrieving program for directing a computer to perform an image retrieving process of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a configuration information generation process of generating similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
a first image feature vector conversion process of converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information;
a second image feature vector conversion process of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information; and
a process of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
37. An image retrieving program for directing a computer to perform an image retrieving process of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a first configuration information prospect selection process of referring to query image feature vector configuration information describing a configuration of an image feature vector of the query image, and outputting a configuration information list of possible feature vector configurations of the image feature vector of the query image;
a second configuration information prospect selection process of referring to images-to-be-retrieved feature vector configuration information describing the configuration of an image feature vector of the image to be retrieved, and outputting a configuration information list of possible feature vector configurations of the image feature vectors of the images to be retrieved;
a configuration information determination process of referring to the configuration information list obtained in said first configuration information prospect selection process and the configuration information list obtained in said second configuration information prospect selection process, and determining similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
a first image feature vector conversion process of converting the image feature vector of the query image into a first image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information;
a second image feature vector conversion process of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information; and
a process of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
38. An image retrieving program for directing a computer to perform an image retrieving process of retrieving an image similar to a predetermined query image from images to be retrieved using an image feature vector, comprising:
a configuration information generation process of generating similarity computation image feature vector configuration information describing a configuration of an image feature vector for use in a similarity computation;
a determination process of determining, according to similarity computation image feature vector configuration information generated in said configuration information generation process or configuration information reselection process, whether or not both of image feature vector of the query image and image feature vectors of the images to be retrieved can be converted;
a configuration information reselection process of selecting similarity computation image feature vector configuration information again when said determination process determines that the image feature vectors cannot be converted;
a first image feature vector conversion process of converting the image feature vector of the query image into a first image feature vector for use in a similarity computation according to the similarity computation image feature vector configuration information when said determination process determines that the image feature vectors can be converted;
a second image feature vector conversion process of converting the image feature vectors of the images to be retrieved into a second image feature vector for use in the similarity computation according to the similarity computation image feature vector configuration information when said determination process determines that the image feature vectors can be converted; and
a process of computing similarity between the first image feature vector and the second image feature vector according to the similarity computation image feature vector configuration information.
39. The image retrieving program according to claim 36, wherein said configuration information generation process refers to query image feature vector configuration information describing a configuration of the image feature vector of the query image and images-to-be-retrieved feature vector configuration information describing a configuration of the image feature vectors of the images to be retrieved, and generates similarity computation image feature vector configuration information.
40. The image retrieving program according to claim 33, wherein when similarity between the first image feature vector and the second image feature vector is computed, normalized similarity can be computed depending on contents of the similarity computation image feature vector configuration information.
41. The image retrieving program according to claim 33, wherein the image feature vector can be generated by obtaining a reduced image from an image of which a feature vector is to be extracted, and performing discrete cosine transform and quantization on the reduced image.
42. The image retrieving program according to claim 33, wherein the similarity computation image feature vector configuration information relates to a configuration of discrete cosine transform coefficients forming an image feature vector.
43. The image retrieving program according to claim 33, wherein the image feature vector is one or more dominant colors of an image, and the similarity computation image feature vector configuration information relates to the dominant color(s).
44. The image retrieving program according to claim 33, wherein the image feature vector indicates a direction of an edge of an image, and the similarity computation image feature vector configuration information relates to a type of the direction of the edge.
45. The image retrieving program according to claim 33, wherein the image feature vector indicates a texture of an image, and the similarity computation image feature vector configuration information relates to a type of the texture.
46. The image retrieving program according to claim 33, wherein the image feature vector indicates a shape of an image, and the similarity computation image feature vector configuration information relates to a type of the shape.
47. The image retrieving program according to claim 33, wherein the image feature vector indicates a motion of an image, and the similarity computation image feature vector configuration information relates to a type of the motion.
48. The image retrieving program according to claim 33, wherein the image feature vector is generated by extracting a feature vector of each rectangular area after dividing an image into rectangular areas, and the similarity computation image feature vector configuration information relates to the rectangular area.