def predictAll(): predictions = [] pictures = glob.glob(testFileDogs) count = len(pictures) i = 0 for each in pictures: image = io.imread(each) image1 = v3(image) image2 = mobile(image) image1 = np.expand_dims(image1, axis=0) image2 = np.expand_dims(image2, axis=0) print("Predicting image " + str(i) + "/" + str(count)) prediction1 = np.argmax(allv3.predict(image1)) prediction2 = np.argmax(allmobile.predict(image2)) if (prediction1 == prediction2): predictions.append([classesAll[prediction1], each.split('/')[-1]]) else: predictions.append([ classesAll[prediction1] + ' or ' + classesAll[prediction2], each.split('/')[-1] ]) i += 1 printall(predictions)
def model1(): predictions = [] pictures = glob.glob(testFile) count = len(pictures) i = 0 for each in pictures: image = io.imread(each) image1 = v3(image) image2 = mobile(image) image1 = np.expand_dims(image1, axis=0) image2 = np.expand_dims(image2, axis=0) print("Predicting image " + str(i) + "/" + str(count)) prediction1 = np.argmax(Cat_Dogv3.predict(image1)) prediction2 = np.argmax(Cat_DogMobileNet.predict(image2)) if (prediction1 == prediction2): if (prediction1 == 0): prediction3 = np.argmax(catv3.predict(image1)) prediction4 = np.argmax(catmobile.predict(image2)) if (prediction3 == prediction4): race = classesCats[prediction3] else: race = [classesCats[prediction3], classesCats[prediction4]] else: prediction3 = np.argmax(dogv3.predict(image1)) prediction4 = np.argmax(dogmobile.predict(image2)) if (prediction3 == prediction4): race = classesDogs[prediction3] else: race = [classesDogs[prediction3], classesDogs[prediction4]] predictions.append( [classesCatDog[prediction1], race, each.split('/')[-1]]) else: predictions.append(['Unsure', each.split('/')[-1]]) i += 1 printall(predictions)
def predictDogandCat(): predictions = [] pictures = glob.glob(testFile) count = len(pictures) i = 0 for each in pictures: image = io.imread(each) image1 = v3(image) image2 = mobile(image) image1 = np.expand_dims(image1, axis=0) image2 = np.expand_dims(image2, axis=0) print("Predicting image " + str(i) + "/" + str(count)) prediction1 = np.argmax(Cat_Dogv3.predict(image1)) prediction2 = np.argmax(Cat_DogMobileNet.predict(image2)) if (prediction1 == prediction2): predictions.append( [classesCatDog[prediction1], each.split('/')[-1]]) else: predictions.append(['Unsure', each.split('/')[-1]]) i += 1 printall(predictions)