def resnet_predict_breed(img_path):
# extract bottleneck features
bottleneck_feature = extract_Resnet50(path_to_tensor(img_path))
# obtain predicted vector
predicted_vector = resnet_model.predict(bottleneck_feature)
# return dog breed that is predicted by the model
return dog_names[np.argmax(predicted_vector)]
def getPredictBreedFromResNet50(imagePath):
# extract bottleneck features
bottleneck_feature = extract_Resnet50(path_to_tensor(imagePath))
# obtain predicted vector
predicted_vector = model_ResNet50.predict(bottleneck_feature)
# return dog breed that is predicted by the model
return dog_names[np.argmax(predicted_vector)]
def Resnet50_dog_breed_prediction(image_path):
# extract bottleneck features
bottleneck_feature = extract_Resnet50(path_to_tensor(image_path))
# obtain predicted vector
prediction_value = dog_breed_model.predict(bottleneck_feature)
# return dog breed name from the list
return dog_names[np.argmax(prediction_value)]
def predict_breed(img_path):
'''
Predicts the dog breed present in the picture
'''
# extract bottleneck features
bottleneck_feature = extract_Resnet50(path_to_tensor(img_path))
# obtain predicted vector
predicted_vector = dogAnalyzer.predict(bottleneck_feature)
# return dog breed that is predicted by the model
name = dog_names[np.argmax(predicted_vector)].split('.')[1]
return name.replace('_', ' ')
def predict_dog_breed(self, img_path):
tensor = self.path_to_tensor(img_path)
print(tensor.shape)
self.bottleneck_features = extract_Resnet50(tensor)
print(self.bottleneck_features.shape)
feat_expand = np.expand_dims(self.bottleneck_features, axis=0)
print(feat_expand.shape)
index = np.argmax(self.Resnet50_model.predict(
self.bottleneck_features))
return self.dog_names[index]
def ResNet50_predict_breed(img_path):
'''
Function to predict the dog breed by using the ResNet50_model on image
IN: img_path - path to image
OUT: predicted dog breed with highest probability
'''
# extract bottleneck features
bottleneck_feature = extract_Resnet50(path_to_tensor(img_path))
# obtain predicted vector
predicted_vector = ResNet50_model.predict(bottleneck_feature)
# return dog breed that is predicted by the model
return dog_names[np.argmax(predicted_vector)]
def getDogPredictFirstThreePlaces(imagePath):
# extract bottleneck features
bottleneck_feature = extract_Resnet50(path_to_tensor(imagePath))
# obtain predicted vector list
predicted_vector = model_ResNet50.predict(bottleneck_feature)[0]
breeds = [
dog_names[index] for index in np.argsort(predicted_vector)[::-1][:3]
]
scores = np.sort(predicted_vector)[::-1][:3]
print('The first three places:')
for index, (breed, score) in enumerate(zip(breeds, scores)):
print('nr: {}, {}, score: {:.6f}'.format(index + 1, breed, score))
showImage(imagePath)
def predict_dog(img_path):
# print(img_path)
img_features = extract_Resnet50(path_to_tensor(img_path))
prediction = np.argmax(Resnet50_model.predict(img_features))
# print(dog_names[prediction])
return dog_names[prediction]