def process(line,d): line = line.strip().split(tab) sample=line[1] image=line[0] featureimages=line[2:] img=Image.open(image) images=imageTransforms.normalizeImage(img) #Arjun's thing features=[] featurecoordinates=[] labels=[] for j in xrange(len(featureimages)): c=getcoordinates.getcoordinates(featureimages[j],sample) for i in xrange(len(c)): featurecoordinates.append(c[i]) del c for j in xrange(len(featurecoordinates)): a=d.get(featurecoordinates[j][5],0) m=max([a,featurecoordinates[j][2]-featurecoordinates[j][0],featurecoordinates[j][3]-featurecoordinates[j][1]]) d[featurecoordinates[j][5]]=m for j in xrange(len(featurecoordinates)): newfeature=runtraininganalysis(images[RGB],featurecoordinates[j]) if newfeature: for i in xrange(len(newfeature)): features.append(newfeature[i]) print("new feature",i) labels.append(featurecoordinates[j][4]+"_"+featurecoordinates[j][5]) labels.append(featurecoordinates[j][4]+"_"+featurecoordinates[j][5]) labels.append(featurecoordinates[j][4]+"_"+featurecoordinates[j][5]) labels.append(featurecoordinates[j][4]+"_"+featurecoordinates[j][5]) return labels,features,d
def process(line): line = line.strip().split(tab) sample=line[1] image=line[0] featureimages=line[2:] img=Image.open(image) normimg=normalize(img) #Arjun's thing features=[] featurecoordinates=[] for j in range(len(featureimages)): inf = featureimages[j].split('.') features.append(inf[1]) c=getcoordinates.getcoordinates(featureimages[j],sample) featurecoordinates.append(c) del c