Image processing method, image processing system and program

1. An image processing method comprising:

extracting a plurality of horizontal segments from given image data;

detecting horizontal segments which correspond to predefined key segments, from among the plurality of horizontal segments;

generating a circumscribed rectangle which is circumscribed to the horizontal segments corresponding to the key segments; and

selecting one form from among form definition structures in which definition information of respective forms is recorded, by utilizing position information, such as vertex positions of the generated circumscribed rectangle;

wherein said key segments are predefined in accordance with criteria which permit the form to be stably selected, and they satisfy a first condition that a thickness of said each key segment is not less than a predetermined threshold value, a second condition that said each key segment doesn't lie in a virtual clear area at a peripheral part of said form, and a third condition that said each key segment lies outside a fold region of said form if any.

2. An image processing method according to claim 1 , wherein if said horizontal segments corresponding to said key segments cannot be detected, the method includes:

detecting horizontal segments which correspond to substitute segments defined for the respective key segments; and

generating said circumscribed rectangle based on information of the substitute segments;

wherein said substitute segments are predefined in accordance with the same criteria as those of said key segments.

3. An image processing method according to claim 2 , wherein if the horizontal segments corresponding to said substitute segments have been detected, it is assumed that horizontal segments exist at positions at which the corresponding key segments ought to exist and which are computed from relative positions of said substitute segments, and a circumscribed rectangle which is circumscribed to the assumed horizontal segments is generated.

4. An image processing method according to any of claim 1 , further comprising recognizing characters included in the image data, by reference to the definition information of the selected form.

5. An image processing method comprising:

extracting a horizontal segment from given image data;

judging if the extracted horizontal segment corresponds to a predefined segment for a rotational direction (“discriminatory segment), so as to determine the image rotational direction;

rotating the image data in the determined rotational direction; and

identifying a form which corresponds to said image data, and if necessary, further recognizing characters included in said image data, by reference to definition information of the form;

wherein said discriminatory segment is predefined in accordance with criteria which permit said form to be stably selected, and it satisfies a first condition that a thickness of said discriminatory segment is not less than a predetermined threshold value, a second condition that said discriminatory segment doesn't lie in a virtual clear area at a peripheral part of said form, a third condition that said discriminatory segment lies outside a fold region of said form if any, and a fourth condition that a segment similar to said discriminatory segment is nonexistent at a position of 180-degree rotational symmetry with respect to a center of said form.

6. An image processing method according to claim 5 , wherein said horizontal segment is extracted from within a partial region of said image data in which said discriminatory segment ought to exist, and a region which is in 180-degree rotational symmetry to the partial region.

7. An image processing method according to claim 6 , comprising:

detecting a horizontal segment which corresponds to a substitute segment defined every rotational direction discriminating segment, if said horizontal segment corresponding to said discriminatory segment cannot be detected; and

judging that said discriminatory segment has been detected, if the horizontal segment corresponding to the substitute segment has been detected.

8. An image processing system comprising:

means for extracting a plurality of horizontal segments from given image data;

means for detecting horizontal segments which correspond to predefined key segments, from among the plurality of horizontal segments;

means for generating a circumscribed rectangle which is circumscribed to the horizontal segments corresponding to the key segments;

form definition structures in which definition information of respective forms is recorded; and

means for selecting one form from among the form definition structures, by utilizing position information, such as vertex positions of the generated circumscribed rectangle;

wherein said key segments are predefined in accordance with criteria which permit the form to be stably selected, and they satisfy a first condition that a thickness of said each key segment is not less than a predetermined threshold value, a second condition that said each key segment doesn't lie in a virtual clear area at a peripheral part of said form, and a third condition that said each key segment lies outside a fold region of said form if any.

9. An image processing system according to claim 8 , comprising:

means for detecting horizontal segments which correspond to substitute segments defined for the respective key segments, if said horizontal segments corresponding to said key segments cannot be detected; and

means for generating said circumscribed rectangle based on information of the substitute segments;

wherein said substitute segments are predefined in accordance with the same criteria as those of said key segments.

10. An image processing system according to claim 9 , wherein if the horizontal segments corresponding to said substitute segments have been detected, it is assumed that horizontal segments exist at positions at which the corresponding key segments ought to exist and which are computed from relative positions of said substitute segments, and a circumscribed rectangle which is circumscribed to the assumed horizontal segments is generated.

11. An image processing system according to claim 8 , further comprising means for recognizing characters included in the image data, by reference to the definition information of the selected form.

12. An image processing system comprising:

means for extracting a horizontal segment from given image data;

means for judging if the extracted horizontal segment corresponds to a predefined segment for a rotational direction (“discriminatory segment”), so as to determine the image rotational direction;

means for rotating the image data in the determined rotational direction; and

means for identifying a form which corresponds to said image data, and if necessary, further recognizing characters included in said image data, by reference to definition information of the form;

wherein said discriminatory segment is predefined in accordance with criteria which permit said form to be stably selected, and it satisfies a first condition that a thickness of said discriminatory segment is not less than a predetermined threshold value, a second condition that said discriminatory segment doesn't lie in a virtual clear area at a peripheral part of said form, a third condition that said discriminatory segment lies outside a fold region of said form if any, and a fourth condition that a segment similar to said discriminatory segment is nonexistent at a position of 180-degree rotational symmetry with respect to a center of said form.

13. An image processing system according to claim 12 , wherein said horizontal segment is extracted from within a partial region of said image data in which said discriminatory segment ought to exist, and a region which is in 180-degree rotational symmetry to the partial region.

14. An image processing system according to claim 13 , comprising:

means for detecting a horizontal segment which corresponds to a substitute segment defined every rotational direction discriminating segment, if said horizontal segment corresponding to said discriminatory segment cannot be detected; and

means for judging that said discriminatory segment has been detected, if the horizontal segment corresponding to the substitute segment has been detected.

15. A computer-executable program stored on a computer readable medium which causes a computer to implement the functions of:

extracting a plurality of horizontal segments from given image data;

detecting horizontal segments which correspond to predefined key segments, from among the plurality of horizontal segments;

generating a circumscribed rectangle which is circumscribed to the horizontal segments corresponding to the key segments;

selecting one form from among form definition structures in which definition information of respective forms is recorded, by utilizing position information, such as vertex positions of the generated circumscribed rectangle; and

recognizing characters included in the image data, by reference to the definition information of the selected form;

wherein said key segments are predefined in accordance with criteria which permit said form to be stably selected, and they satisfy a first condition that a thickness of said each key segment is not less than a predetermined threshold value, a second condition that said each key segment doesn't lie in a virtual clear area at a peripheral part of said form, and a third condition that said each key segment lies outside a fold region of said form if any.

16. A computer-executable program stored on a computer readable medium which causes a computer to implement the functions of:

extracting a horizontal segment from given image data;

judging if the extracted horizontal segment corresponds to a predefined segment for a rotational direction (“discriminatory segment”), so as to determine the image rotational direction;

rotating the image data in the determined rotational direction; and

identifying a form which corresponds to said image data, and recognizing characters included in said image data, by reference to definition information of the form;

wherein said discriminatory segment is predefined in accordance with criteria which permit said form to be stably selected, and it satisfies a first condition that a thickness of said discriminatory segment is not less than a predetermined threshold value, a second condition that said discriminatory segment doesn't lie in a virtual clear area at a peripheral part of said form, a third condition that said discriminatory segment lies outside a fold region of said form if any, and a fourth condition that a segment similar to said discriminatory segment is nonexistent at a position of 180-degree rotational symmetry with respect to a center of said form.