Fast regular shaped pattern searching
View Patent ↗An accumulation method for fast pattern search can accurately locate regular shaped patterns of interest. The method can be used for invariant search to match patterns of interest in images where the searched pattern varies in size or orientation or aspect ratio, when pattern appearance is degraded, and even when the pattern is partially occluded, where the searched image is large, multidimensional, or very high resolution, or where the pattern size is large. The computation requirement is independent of the size of the pattern region.
1. A fast pattern search method comprising the steps of:
a) Input at least one image;
b) Input a pattern template specification;
c) Using the at least one input image and the pattern template specification, perform an accumulation transformation having at least one accumulation image output;
d) Perform pattern kernel construction using the at least one accumulation image output to generate at least one pattern kernel result image;
e) Perform a rotation adjustment generation step to guide accumulation transformation having a set of accumulation image outputs;
f) Perform rotation invariant pattern search using the accumulation image outputs to create a rotation invariant pattern search result output.
2. A fast pattern search method comprising the steps of:
a) Input at least one image;
b) Input a pattern template specification;
c) Using the at least one input image and the pattern template specification, perform an accumulation transformation having at least one accumulation image output;
d) Perform pattern kernel construction using the at least one accumulation image output to generate at least one pattern kernel result image;
e) Perform a scale adjustment generation step to guide pattern kernel construction having a set of pattern kernel result image outputs;
f) Perform scale invariant pattern search using the at least one accumulation image and pattern kernel result image outputs to create an invariant pattern search result output.
3. A fast pattern search method comprising the steps of:
a) Input at least one image;
b) Input a pattern template specification;
c) Using the at least one input image and the pattern template specification, perform an accumulation transformation having at least one accumulation image output;
d) Perform pattern kernel construction using the at least one accumulation image output to generate at least one pattern kernel result image;
e) Perform an aspect ratio adjustment generation step to guide pattern kernel construction having a set of pattern kernel result image outputs;
f) Perform aspect ratio invariant pattern search using the at least one accumulation image and pattern kernel result image outputs to create an aspect ratio invariant pattern search result output.
4. A fast pattern search method comprising the steps of:
a) Input at least one image;
b) Input a pattern template specification;
c) Using the at least one input image and the pattern template specification, perform an accumulation transformation having at least one accumulation image output;
d) Perform pattern kernel construction using the at least one accumulation image output to generate at least one pattern kernel result image;
e) Perform a scale and aspect ratio adjustment generation step and a rotation adjustment generation step having a set of accumulation image outputs and a set of pattern kernel result image outputs;
f) Perform an invariant pattern search selected from the set consisting of rotation, scale, and aspect ratio using the set of accumulation image outputs and the set of pattern kernel result image outputs.
5. A fast pattern search method comprising the steps of:
a) Input at least one image;
b) Input a pattern template specification wherein the pattern template specification includes pattern approximation wherein pattern approximation includes non-uniform gray patterns;
c) Using the at least one input image and the pattern template specification, perform an accumulation transformation having at least one accumulation image output;
d) Perform pattern kernel construction using the at least one accumulation image output to generate at least one pattern kernel result image;
e) Perform a pattern search using the at least one accumulation image and the at least one pattern kernel result image to create a pattern search result output.
6. A fast pattern search method comprising the steps of:
a) Input at least one image;
b) Input a pattern template specification;
c) Using the at least one input image and the pattern template specification, perform an accumulation transformation having at least one accumulation image output;
d) Perform pattern kernel construction using the at least one accumulation image output to generate at least one pattern kernel result image wherein the pattern kernel construction step includes a composite positive and negative pattern construction
e) Perform a first positive- pattern- only search to create good match candidate outputs;
f) Further process the good match candidate outputs using a composite positive and negative pattern search to create a pattern search result output.