Pairwise Matching

runGraphMatchBatch

Run Graph Matching on a batch of images.

pMatch = runGraphMatchBatch(datapath,viewList,'all', 0, wEdge=0);

Inputs

• datapath: filepath of image class
• viewList: list of hypercols files for class
• pairingMode: type of pairing mode for matches (all, loop, neighbor)
• pairingParam: the radius of the neighborhood if pairing mode is loop or neighbor
• wEdge: type of linear/graph matching (0, 1, 2, ect)

Outputs

• pairMatches: Numpy array of pairwiseMatches class, shape is (len(viewList), len(viewList))

Other Method Calls

• graphMatch

graphMatch

Run Graph Matching on two images.

matchData = graphMatch(viewPair, wEdge)

Inputs

• viewPair: tuple of two hypercols filepaths
• (optional)methodGM: Method for solving graph matching (sm, rw)
• (optional)methodDisc: Method for discretization (greedy, hungry)
• (optional)wEdge: Weigth of rigidicity
• (optional)wAngle: no idea tbh
• (optional)kNNInit: number of candidate correpondences for each feature
• (optional)nMaxInit: total number of candidate correpondences after optimization
• (optional)thScore: threshold of matching scores below which the match is ignored
• (optional)thRatio: no idea, threshold ratio ?
• (optional)thDist: no idea, threshold distance ?

Outputs

• mData: a pairwiseMatches class

Other Method Calls

• initMatch
• greedyMatch

initMatch

Performs initial Matching simScores, matchInds = initMatch(viewPair,kNNInit,nMaxInit,thScore,thRatio)

Inputs

• viewPair: tuple of two hypercols filpaths
• kNNInit: k value for k nearest neighbors
• nMaxInit:
• threshScore:
• threshRatio:

Outputs

• simScores: Numpy array of similarity scores
• matchInds: Numpy array match indices

Other Method Calls

greedyMatch

Performs Greedy Matching X = greedyMatch(matchInfo, Xraw)

Inputs

• match: numpy array of matches
• score: Xraw, similarity scores, 1D, horizontal
• (optional)nMax: kinda wack, convergence condition?

Outputs

• flag: numpy array of greedy matches, (score.shape[0], 1)

Other Method Calls