GitHub - RajatBhageria/EdgeMatching: Edge Matching for Object Recognition of Pills Based on Canny Edge Detector

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 28 Commits
library		library
query		query
README		README
Report_Pill Object Classification Using Chamfer Edge Matching.pdf		Report_Pill Object Classification Using Chamfer Edge Matching.pdf
Test_script.m		Test_script.m
cannyEdge.m		cannyEdge.m
computeCostOfDifference.m		computeCostOfDifference.m
correspondPixels.m		correspondPixels.m
correspondPixels.mexa64		correspondPixels.mexa64
correspondPixels.mexmaci64		correspondPixels.mexmaci64
correspondPixels.mexw64		correspondPixels.mexw64
createLibraryForEdgeMapping.m		createLibraryForEdgeMapping.m
cropEdgeMap.m		cropEdgeMap.m
edgeEval.m		edgeEval.m
edgeLink.m		edgeLink.m
edgeMatching.m		edgeMatching.m
findDerivatives.m		findDerivatives.m
nonMaxSup.m		nonMaxSup.m
rajat_kashishg_EdgeMatching.zip		rajat_kashishg_EdgeMatching.zip
visDerivatives.m		visDerivatives.m

Repository files navigation

The main class for our edge matching algorithm is edgeMatching.m

It takes in two one parameter, a nxmx3 query image. Note that right now our algorithm is based on the images found in the /library folder and thus only works with images in our /query folder. 

Simply call edgeMatching(query) with a nxmx3 query image from our /query class and our matching algorithm should tell you whether the image is "Accurpril", "Vicodin", "Advil C&S", or "Ibuprofen"