app.py

import streamlit as st
import tensorflow as tf
import tensorflow_hub as hub
from PIL import Image
import numpy as np
import cv2
import io
import os  

# Load the MobileNet model from TensorFlow Hub
mobilenet_url = "https://tfhub.dev/google/tf2-preview/mobilenet_v2/classification/4"
mobilenet_model = hub.load(mobilenet_url)

# Create a function to perform image classification
def classify_image(image):
    # Preprocess the image
    image = image.resize((224, 224))
    image = np.array(image) / 255.0  # Normalize the pixel values
    image = tf.image.convert_image_dtype(image, tf.float32)
    image = tf.image.resize_with_pad(image, 224, 224)
    image = tf.expand_dims(image, axis=0)

    # Perform classification
    predictions = mobilenet_model(image)
    return predictions

# Function to perform basic deepfake detection (for educational purposes)
def detect_deepfake(image_path):
    image = cv2.imread(image_path)
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')
    faces = face_cascade.detectMultiScale(gray_image, scaleFactor=1.1, minNeighbors=5)

    if len(faces) > 0:
        return 0.3  # A face was detected, indicating it may be a real image
    else:
        return 0.9  # No face detected; it may be a deepfake

# Streamlit UI
st.title("Truecaptions - AI-Generated Image Detection")
uploaded_image = st.file_uploader("Upload an image", type=["jpg", "png", "jpeg"])

if uploaded_image is not None:
    image = Image.open(uploaded_image)
    st.image(image, caption="Uploaded Image", use_column_width=True)

    if st.button("Classify"):
        st.write("Classifying...")
        predictions = classify_image(image)

        st.write("Image Classification Results:")
        top_classes = np.argsort(predictions)[0, ::-1][:5]  # Top 5 classes
        for i, class_idx in enumerate(top_classes):
            st.write(f"Class {class_idx}: {predictions[0, class_idx]:.4f}")

        st.write("Detecting deepfake...")

        # Save the uploaded image to a temporary file and get its path
        temp_image_path = "temp_image.jpg"
        image.save(temp_image_path)

        deepfake_score = detect_deepfake(temp_image_path)

        # Remove the temporary image file
        os.remove(temp_image_path)

        if deepfake_score > 0.5:
            st.write("This image is likely AI-generated (deepfake).")
        else:
            st.write("This image appears to be real.")

st.text("Upload an image and click the 'Classify' button to see image classification and deepfake detection results.")