app.py

import tensorflow.keras as keras
import extract_bottleneck_features
import cv2      
import gradio as gr
import numpy as np      
from glob import glob
from keras.preprocessing import image           
InceptionV3_model = keras.models.load_model("weights.best.InceptionV3.hdf5",)

#dog_names = [item[20:-1] for item in sorted(glob("dogImages/train/*/"))]

dog_names= ['Affenpinscher', 'Afghan_hound', 'Airedale_terrier', 'Akita', 'Alaskan_malamute', 'American_eskimo_dog', 'American_foxhound', 'American_staffordshire_terrier', 'American_water_spaniel', 'Anatolian_shepherd_dog', 'Australian_cattle_dog', 'Australian_shepherd', 'Australian_terrier', 'Basenji', 'Basset_hound', 'Beagle', 'Bearded_collie', 'Beauceron', 'Bedlington_terrier', 'Belgian_malinois', 'Belgian_sheepdog', 'Belgian_tervuren', 'Bernese_mountain_dog', 'Bichon_frise', 'Black_and_tan_coonhound', 'Black_russian_terrier', 'Bloodhound', 'Bluetick_coonhound', 'Border_collie', 'Border_terrier', 'Borzoi', 'Boston_terrier', 'Bouvier_des_flandres', 'Boxer', 'Boykin_spaniel', 'Briard', 'Brittany', 'Brussels_griffon', 'Bull_terrier', 'Bulldog', 'Bullmastiff', 'Cairn_terrier', 'Canaan_dog', 'Cane_corso', 'Cardigan_welsh_corgi', 'Cavalier_king_charles_spaniel', 'Chesapeake_bay_retriever', 'Chihuahua', 'Chinese_crested', 'Chinese_shar-pei', 'Chow_chow', 'Clumber_spaniel', 'Cocker_spaniel', 'Collie', 'Curly-coated_retriever', 'Dachshund', 'Dalmatian', 'Dandie_dinmont_terrier', 'Doberman_pinscher', 'Dogue_de_bordeaux', 'English_cocker_spaniel', 'English_setter', 'English_springer_spaniel', 'English_toy_spaniel', 'Entlebucher_mountain_dog', 'Field_spaniel', 'Finnish_spitz', 'Flat-coated_retriever', 'French_bulldog', 'German_pinscher', 'German_shepherd_dog', 'German_shorthaired_pointer', 'German_wirehaired_pointer', 'Giant_schnauzer', 'Glen_of_imaal_terrier', 'Golden_retriever', 'Gordon_setter', 'Great_dane', 'Great_pyrenees', 'Greater_swiss_mountain_dog', 'Greyhound', 'Havanese', 'Ibizan_hound', 'Icelandic_sheepdog', 'Irish_red_and_white_setter', 'Irish_setter', 'Irish_terrier', 'Irish_water_spaniel', 'Irish_wolfhound', 'Italian_greyhound', 'Japanese_chin', 'Keeshond', 'Kerry_blue_terrier', 'Komondor', 'Kuvasz', 'Labrador_retriever', 'Lakeland_terrier', 'Leonberger', 'Lhasa_apso', 'Lowchen', 'Maltese', 'Manchester_terrier', 'Mastiff', 'Miniature_schnauzer', 'Neapolitan_mastiff', 'Newfoundland', 'Norfolk_terrier', 'Norwegian_buhund', 'Norwegian_elkhound', 'Norwegian_lundehund', 'Norwich_terrier', 'Nova_scotia_duck_tolling_retriever', 'Old_english_sheepdog', 'Otterhound', 'Papillon', 'Parson_russell_terrier', 'Pekingese', 'Pembroke_welsh_corgi', 'Petit_basset_griffon_vendeen', 'Pharaoh_hound', 'Plott', 'Pointer', 'Pomeranian', 'Poodle', 'Portuguese_water_dog', 'Saint_bernard', 'Silky_terrier', 'Smooth_fox_terrier', 'Tibetan_mastiff', 'Welsh_springer_spaniel', 'Wirehaired_pointing_griffon', 'Xoloitzcuintli', 'Yorkshire_terrier']

labels =  dog_names

def extract_InceptionV3(tensor):
	from keras.applications.inception_v3 import InceptionV3, preprocess_input
	return InceptionV3(weights='imagenet', include_top=False).predict(preprocess_input(tensor))


def extract_Resnet50(tensor):
	from tensorflow.keras.applications.resnet50 import ResNet50, preprocess_input
	return ResNet50(weights='imagenet', include_top=False).predict(preprocess_input(tensor))


###########################################

from tensorflow.keras.applications.resnet50 import preprocess_input

######################################

import tensorflow as tf
from keras.preprocessing import image                  
from tqdm import tqdm

######################################

from tensorflow.keras.applications.resnet50 import ResNet50
# define ResNet50 model
ResNet50_model = ResNet50(weights='imagenet')

from keras.preprocessing import image                  
from tqdm import tqdm



from tensorflow.keras.applications.resnet50 import preprocess_input

def ResNet50_predict_labels(img):
    # returns prediction vector for image located at img_path
    img = np.expand_dims(img, axis=0)
    img = preprocess_input((img))
    return np.argmax(ResNet50_model.predict(img)) 


def path_to_tensor(img_path):
    # loads RGB image as PIL.Image.Image type
    #img = image.load_img(img_path, target_size=(224, 224))
    # convert PIL.Image.Image type to 3D tensor with shape (224, 224, 3)
    #x = image.img_to_array(img)
    # convert 3D tensor to 4D tensor with shape (1, 224, 224, 3) and return 4D tensor
    return np.expand_dims(img_path, axis=0)



# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_alt.xml')



def face_detector(image):
    """
    returns "True" if face is detected in image stored at image

    """
    
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray)
    if len(faces) > 0:
        return "Number of human faces found in this image: {}". format(len(faces))
    else:
        return "There are no human faces in this image"




def InceptionV3_prediction_breed(img_path):
    """
    Return: dog breed that is predicted by the model
    input: image
    """
    
    # extract bottleneck features
    bottleneck_feature = extract_InceptionV3(path_to_tensor(img_path))
    # obtain predicted vector
    predicted_vector = InceptionV3_model.predict(bottleneck_feature)
    # return dog breed that is predicted by the model
    return dog_names[np.argmax(predicted_vector)].split('.')[-1]



def dog_detector(img):
    """
    input: uploaded image by user
    return: "True" if a dog is detected in the image stored at img
    """
    
    prediction = ResNet50_predict_labels(img)
    return ((prediction <= 268) & (prediction >= 151)) 

def identify_dog_app(img):
    """This function predicts the breed of the human or dog"
    
    input: uploaded image by user
    Return: dog or human, and breed of the uploaded image
    """
    
    breed = InceptionV3_prediction_breed(img)
    if dog_detector(img):
        return("This looks like a dog and its breed is:"),"{}".format(breed)
    elif face_detector(img):
        return("This looks like a human but might be classified as a dog of the following breed:"),"{}".format(breed)
    else:
        return("I have no idea what this might be. Please upload another image!"), ("Not applicable")




image = gr.inputs.Image(shape=(224, 224),  label="Image")
label = gr.outputs.Label(num_top_classes=1)    

iface = gr.Interface(   
    fn=identify_dog_app,
    inputs=image,
    outputs=[gr.outputs.Label(label="Human or Dog?"), gr.outputs.Label(label="Breed:")],  
    title="Human or dog Identification - Breed Classification",
    #description ="Please find the jypyter notebook on ___",
    article = 
    '<b><span style="color: #ff9900;">Acknowledgement:</span></b><br/>'
    +'<p><span style="color: #ff9900;">I would like to express my special thanks of gratitude'
    +'to Misk &amp; Sdaia for giving me the opportunity to enrol in "Data Scientist" Udacity nanodegree,'
    +'&nbsp;as well as to my mentor Mr. Haroon who was of great help during my learning journey.</span></p>'
    +'<p><span style="color: #ff9900;">This is my capstone project and herewith I finish this ND.</span></p>',

    theme="dark-huggingface"

)

iface.launch(share=False)