def __init__(self, dataset, k, disks, circumferences):
self.dataset = dataset
self.k = k
self.disks = disks
self.circumferences = circumferences
self.sim = SimilaritySearch()
self.train()
class FindK():
def __init__(self, dataset, k, disks, circumferences):
self.dataset = dataset
self.k = k
self.disks = disks
self.circumferences = circumferences
self.sim = SimilaritySearch()
self.train()
def train(self):
for i in range(0, self.dataset.getTrainingSize()):
hist = self.dataset.trainingData[i]
label = self.dataset.trainingLabels[i]
self.sim.addHistogram(hist, label)
self.sim.train()
def getTopK(self, neighbours, dist, labels, k):
results = []
resultsdists = []
cont = 0
neighbours = neighbours.flatten()
dist = dist.flatten()
while cont < len(neighbours) and k > 0:
pos = neighbours[cont]
item = labels[pos]
item.setDistance(dist[cont])
if not item in results and dist[cont] > 0:
results.append(item)
resultsdists.append(dist[cont])
k -= 1
cont += 1
return results
def find(self, img, ht):
print("Searching...")
speciesImage = SpeciesImage(inputImg=img,
disks=self.disks,
circumferences=self.circumferences)
neighbours, dist = self.sim.findk(speciesImage.getHistogram(ht))
results = self.getTopK(neighbours, dist, self.dataset.labels, self.k)
return results
