app.py

# -*- coding: utf-8 -*-

# %%capture
# #Use capture to not show the output of installing the libraries!

#model_multi = tf.keras.models.load_model("densenet")

# define the labels for the multi-label classification model
#labels_multi = {0: 'healthy', 1: 'mild', 2: 'moderate'}


#model = tf.keras.models.load_model('/content/drive/MyDrive/project_image_2023_NO/saved_models/saved_model/densenet')
#labels = ['Healthy', 'Patient']
#labels = {0: 'healthy', 1: 'patient'}

import gradio as gr
import requests
import torch
import torch.nn as nn
from PIL import Image
from torchvision.models import resnet50
from torchvision.transforms import functional as F
import numpy as np
import tensorflow as tf
from transformers import pipeline
from tensorflow.keras.preprocessing import image as image_utils
from tensorflow.keras.applications import densenet, efficientnet


import tensorflow as tf
import gradio as gr

# load the CNN binary classification model
model_cnn = tf.keras.models.load_model("CNN_binary")

# define the labels for the binary classification model
labels_cnn = {0: 'Healthy', 1: 'Patients'}

# load the EfficientNet binary classification model
model_efn = tf.keras.models.load_model("efficientNet_binary")

# define the labels for the binary classification model
labels_efn = {0: 'Healthy', 1: 'Patients'}

#def classify_cnn(inp):
    #inp = inp.reshape((-1, 224, 224, 3))
    #inp = tf.keras.applications.densenet.preprocess_input(inp)
    #prediction = model_cnn.predict(inp)
    #confidences = {labels_cnn[i]: float(prediction[0][i]) for i in range(2)}
    #return confidences

#def classify_efn(inp):
    #inp = inp.reshape((-1, 224, 224, 3))
    #inp = tf.keras.applications.efficientnet.preprocess_input(inp)
    #prediction = model_efn.predict(inp)
    #confidences = {labels_efn[i]: float(prediction[0][i]) for i in range(2)}
    #return confidences

def classify_cnn(inp):
    inp = inp.reshape((-1, 224, 224, 3))
    inp = tf.keras.applications.densenet.preprocess_input(inp)
    prediction = model_cnn.predict(inp)
    class_index = np.argmax(prediction, axis=-1)[0]
    return labels_cnn[class_index]

def classify_efn(inp):
    inp = inp.reshape((-1, 224, 224, 3))
    inp = tf.keras.applications.efficientnet.preprocess_input(inp)
    prediction = model_efn.predict(inp)
    class_index = np.argmax(prediction, axis=-1)[0]
    return labels_efn[class_index]


binary_interface_cnn = gr.Interface(fn=classify_cnn, 
                                inputs=gr.Image(shape=(224, 224)),
                                #outputs=gr.Label(num_top_classes=2),
                                outputs=gr.outputs.Textbox(),
                                title="Binary Image Classification",
                                description="Classify an image as healthy or patient using custom CNN.",
                                examples=[['3310277.png'],['371129.png']]
                               )


binary_interface_efn = gr.Interface(fn=classify_efn, 
                               inputs=gr.Image(shape=(224, 224)),
                               #outputs=gr.Label(num_top_classes=2),
                               outputs=gr.outputs.Textbox(),
                               title="Binary Image Classification",
                               description="Classify an image as healthy or patient using EfficientNet.",
                               examples=[['3310277.png'],['371129.png']]
                              )





demo = gr.TabbedInterface([binary_interface_cnn, binary_interface_efn], ["Custom CNN", "CNN"])

demo.launch()