Pattern registration
A system, method, and computer program product for automatic machine-implemented pattern registration, particularly suited to small sensor systems which include a reduced image capture area size relative to the pattern being sensed.
1. A machine-implemented method for processing an input set of pattern structures, the input set of pattern structures derived from an N number of pattern representations from one or more pattern sources, N an integer greater than 1, comprising executing on a processor:
a) setting, using the processor, a link quality metric;
b) generating, using the processor, a link list from an evaluation of an M number of unique pair-wise match comparisons of pattern structure pairs from the input set of pattern structures, M an integer less than or equal to (N*(N−1))/2;
c) producing, using the processor and responsive to said link list, a link list quality metric; and
d) identifying, using the processor responsive to said link list, a linked set of pattern structures from the input set of pattern structures when said link list quality metric has a predetermined relationship to said link list quality metric.
2. The method of claim 1 wherein said link quality metric includes a quantification of a number of unique pattern structures from the input set of pattern structures to be included in said linked set of pattern structures.
3. The method of claim 2 wherein said link list quality metric includes a threshold of a minimum number of said M number of unique pair-wise match comparisons which include a match true state and wherein said predetermined relationship includes at least equaling said threshold.
4. The method of claim 1 wherein said linked set of pattern structures consists essentially of pattern structures from the input set of pattern structures that match at least one other pattern structure from the input set of pattern structures.
5. The method of claim 1 further comprising a trusted memory store as part of a secure pattern structure registration system and recording said linked set of pattern structures into said trusted memory store.
6. The method of claim 4 further comprising a trusted memory store as part of a secure pattern structure registration system and recording said linked set of pattern structures into said trusted memory store.
7. A machine-implemented method for identifying a linked set of pattern structures from an input set of pattern structures, the input set of pattern structures derived from an N number of pattern representations from one or more pattern sources, N an integer greater than 1 with the linked set of pattern structures including an identification of each of a first pattern structure from the input set of pattern structures matching at least one of a second pattern structure different from the first pattern structure, comprising executing on a processor:
a) setting, using the processor, a link quality metric;
b) performing, using the processor, a pair-wise match process of a unique two-combination of a first particular pattern structure from the input set of pattern structures against a second particular pattern structure from the input set of pattern structures, the second particular pattern structure different from said first particular pattern structure, each said pair-wise match process associating, for each said unique two-combination, a MATCH value for a set of MATCH values wherein each said MATCH value includes a match true indication or a match false indication;
c) associating, using the processor when said MATCH value includes said match true indication for said particular pattern structures of each said unique two-combination, said particular pattern structures in a link list;
d) performing for each step c), using the processor responsive to said link list, a link quality score;
e) comparing, using the processor, said link quality score against said link quality metric to produce a link list evaluation metric;
f) repeating step b)-step f), using the processor, when said link list evaluation metric includes a first predetermined metric;
g) including, using the processor, said particular pattern structures associated with said link list in the linked set of pattern structures when said link list evaluation metric includes a second predetermined metric.
8. The method of claim 7 wherein said link list quality metric includes a quantification of a number of unique pattern structures from the input set of pattern structures to be included in the linked set of pattern structures.
9. The method of claim 8 wherein said link quality metric includes a threshold of a minimum number of an M number of unique pair-wise match comparisons, M an integer less than or equal to (N*(N−1))/2, which include a match true state and wherein said predetermined relationship includes at least equaling said threshold.
10. The method of claim 7 wherein the linked set of pattern structures consists essentially of pattern structures from the input set of pattern structures that match at least one other pattern structure from the input set of pattern structures.
11. The method of claim 7 further comprising a trusted memory store as part of a secure pattern structure registration system and recording the linked set of pattern structures into said trusted memory store.
12. The method of claim 10 further comprising a trusted memory store as part of a secure pattern structure registration system and recording the linked set of pattern structures into said trusted memory store.