Feature recognition using loose gray scale template matching
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06K-009/62
G06K-009/64
출원번호
US-0745623
(2000-12-26)
발명자
/ 주소
Loce, Robert P.
Handley, John C.
Cuciurean-Zapan, Clara
출원인 / 주소
Xerox Corporation
인용정보
피인용 횟수 :
14인용 특허 :
21
초록▼
What is presented is a method for feature recognition using loose-gray-scale template matching including at least an initial point for locating within the received image a plurality of pixel points in gray-scale surrounding an initial point. The method has the steps of first locating the initial poi
What is presented is a method for feature recognition using loose-gray-scale template matching including at least an initial point for locating within the received image a plurality of pixel points in gray-scale surrounding an initial point. The method has the steps of first locating the initial point and the plurality of pixel points to define feature boundaries and then generating a looseness interval about the located initial point with template information being associated therewith. The next step involves determining which one of a plurality of templates for fitting fits within a threshold looseness interval, and then outputting a signal associated with that recognized feature.
대표청구항▼
1. A method for feature recognition using a method for matching a plurality of templates with a received image comprising:(a) receiving an image comprised of gray-scale image data;(b) generating a two-dimensional window of the received gray-scale image data, the two-dimensional window having a targe
1. A method for feature recognition using a method for matching a plurality of templates with a received image comprising:(a) receiving an image comprised of gray-scale image data;(b) generating a two-dimensional window of the received gray-scale image data, the two-dimensional window having a target gray-scale pixel and a plurality of surrounding gray-scale pixels, each gray-scale pixel being associated with a pixel of received gray-scale image data in the two-dimensional window;(c) determining a plurality of looseness intervals for one template of a plurality of templates, each looseness interval being a difference between a gray-scale pixel associated with the two-dimensional window and a template gray scale pixel from the template and corresponding to the pixel location in the two-dimensional window;(d) determining a template looseness interval value for the template based upon the determined plurality of looseness intervals;(e) comparing the determined template looseness interval value to a threshold looseness interval value, the threshold looseness interval value being a maximum allowable value for the determined template looseness interval value that indicates a loosely matched template;(f) determining, based upon the comparison between the determined template looseness interval value and the threshold looseness interval, which template of the plurality of templates loosely matches the two-dimensional window of received gray-scale image data, the loosely matched template being a template wherein the determined template looseness interval value associated therewith is equal to a non-zero value and the threshold looseness interval value is equal to a non-zero value, an exactly matched template being a template wherein the associated determined looseness interval value is equal to a zero value and the threshold looseness interval value is equal to a zero value; and(g) outputting an identifier associated with the loosely matched template such that the identifier indicates a recognized image feature. 2. The method as in claim 1, wherein the recognized image feature may a feature from a group of features including serifs, corners, vertical strokes, horizontal strokes, diagonal strokes, glyphs, platen cover marks, ink traps, paper edges, registration marks, binding holes and marks, or presence of adjacent color edges. 3. The method as in claim 1, further comprising:(h) determining from a plurality of identifiers a recognized global image feature. 4. The method as in claim 3, wherein the recognized global image feature may a global feature from a group of global features including financial documents, secure documents, bonds, stamps, or passports. 5. The method as in claim 1, further comprising:(h) increasing the contrast of the recognized feature based upon the outputted identifier. 6. The method as in claim 1, further comprising:(h) darkening the recognized feature based upon the outputted identifier. 7. The method as in claim 1, further comprising:(h) thinning the recognized feature based upon the outputted identifier. 8. The method as in claim 1, wherein the recognized image feature is an adjacent colored edged. 9. The method as in claim 1, wherein the determined plurality of looseness intervals are determined from various color separations. 10. The method as claimed in claim 1, wherein the determined looseness interval value for the template is the average looseness interval value determined from the plurality of determined looseness intervals. 11. The method as in claim 1, further comprising:(h) using, when more than one template loosely matches the two-dimensional window of received gray-scale image data, an arbitration process to select the most preferred loosely matched template. 12. The method as in claim 1, further comprising:(h) creating an image plane being composed of the outputted identifiers. 13. A method for feature recognition using a method for matching a plurality of templates with a received image comprisi ng:(a) receiving an image comprised of gray-scale image data;(b) generating a two-dimensional window of the received gray-scale image data, the two-dimensional window having a target gray-scale pixel and a plurality of surrounding gray-scale pixels, each gray-scale pixel being associated with a pixel of received gray-scale image data in the two-dimensional window;(c) determining a plurality of looseness intervals for one template from a plurality of templates, each looseness interval being a difference between a gray-scale pixel associated with the two-dimensional window and a template gray scale pixel from the template and corresponding to the pixel location in the two-dimensional window;(d) comparing each non-zero looseness interval of the plurality of looseness intervals with a threshold looseness interval to determine a looseness interval number for the template, the looseness interval number being equal to how many non-zero looseness intervals within the set of looseness intervals are less than the threshold looseness interval;(e) selecting the template of the plurality of templates having a greatest associated looseness interval number as the template that loosely matches the two-dimensional window of received gray-scale image data; and(f) outputting an identifier associated with the loosely matched template such that the identifier indicates a recognized image feature. 14. The method as in claim 13, wherein the recognized image feature may a feature from a group of features including serifs, corners, vertical strokes, horizontal strokes, diagonal strokes, glyphs, platen cover marks, ink traps, paper edges, registration marks, binding holes and marks, or presence of adjacent color edges. 15. The method as in claim 13, further comprising:(g) increasing the contrast of the recognized feature based upon the outputted identifier. 16. The method as in claim 13, further comprising:(g) darkening the recognized feature based upon the outputted identifier. 17. The method as in claim 13, further comprising:(g) thinning the recognized feature based upon the outputted identifier. 18. The method as in claim 13, wherein the recognized image feature is an adjacent colored edged. 19. The method as in claim 13, further comprising:(g) using, when more than one template loosely matches the two-dimensional window of received gray-scale image data, an arbitration process to select the most preferred loosely matched template. 20. The method as in claim 13, further comprising:(g) creating an image plane being composed of the outputted identifiers.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (21)
Alvarez Jose R. ; Herz William S., Adaptive dynamic aperture correction.
Curry Douglas N. (Menlo Park CA) St. John Robert P. (Santa Clara CA) Filshtinsky Sofia (Saratoga CA), Enhanced fidelity reproduction of images by hierarchical template matching.
Petrick Bruce E. (Lake Oswego OR) Wingfield Perry E. (Tigard OR), Method and apparatus for input picture enhancement by removal of undersired dots and voids.
Lin Ying-Wei ; Loce Robert P. ; Branciforte Michael, Method and apparatus for the resolution enhancement of gray-scale images that include text and line art.
Driscoll ; Jr. Edward C. (Portola Valley CA) Martin Craig O. (Menlo Park CA) Ruby Kenneth (Florence OR) Russell James J. (Mountain View CA) Watson John G. (Menlo Park CA), Method and apparatus for verifying identity using image correlation.
Loce Robert P. (Rochester NY) Cianciosi Michael S. (Rochester NY), Method for design and implementation of an image resolution enhancement system that employs statistically generated look.
Coonan Bonnie R. (Rochester NY) Frumusa Anthony M. (Penfield NY) Nacman Aron (Penfield NY) Tse Francis K. (Rochester NY) Davidson Michael L. (Rochester NY), Multifunctional apparatus for appearance tuning and resolution reconstruction of digital images.
Loce Robert P. (Rochester NY) Cianciosi Michael S. (Rochester NY) Klassen R. Victor (Webster NY), Non-integer image resolution conversion using statistically generated look-up tables.
Tse, Francis Kapo; Nagarajan, Ramesh; Cassidy, John Christopher; Lee, Chia-Hao; Wu, John W., Methods for background and noise suppression in binary to grayscale image conversion.
Beauchaine, Robert J.; Chabreck, Thomas E.; Howells, Samuel C.; Hubbard, John J.; Klatchko, Asher; Pirogovsky, Peter; Teitzel, Robin L., Optical proximity correction in raster scan printing based on corner matching templates.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.