Update app.py

app.py

@@ -1,54 +1,44 @@
-import streamlit as st
-import cv2
-import os
-import tempfile
+import gradio as gr
 import numpy as np
+import cv2
 from prediction import Prediction
-from PIL import Image
-
-st.set_page_config(page_title="Deepfake Detection", layout="wide")
-st.title("🔍 Deepfake Video Detector with Grad-CAM Overlay")
 
-st.markdown(
-    "Upload a video file below to analyze it for deepfakes. The model will detect faces, "
-    "analyze the content, and display the result with a Grad-CAM overlay for interpretability."
+# Initialize the Prediction class
+predictor = Prediction()
+
+def inference(video):
+    """
+    Gradio-compatible inference function.
+
+    Args:
+        video (str): Path to the uploaded video file
+
+    Returns:
+        tuple: (Prediction string, Grad-CAM image, Classification details)
+    """
+    prediction, gradcam_image, classification_details = predictor.predict(video)
+
+    # Convert Grad-CAM image to RGB for display in Gradio
+    if gradcam_image is not None:
+        gradcam_image = cv2.cvtColor(gradcam_image, cv2.COLOR_BGR2RGB)
+    else:
+        gradcam_image = np.zeros((256, 256, 3), dtype=np.uint8)  # fallback image
+
+    return prediction, gradcam_image, classification_details
+
+# Define Gradio interface
+demo = gr.Interface(
+    fn=inference,
+    inputs=gr.Video(label="Upload a video"),
+    outputs=[
+        gr.Textbox(label="Prediction"),
+        gr.Image(label="Grad-CAM Visualization"),
+        gr.Label(label="Detailed Classification"),
+    ],
+    title="Deepfake Detection with Grad-CAM",
+    description="Upload a video and detect whether it is real or a deepfake. Grad-CAM will highlight the most influential region.",
+    theme="default",
 )
 
-# File uploader
-uploaded_file = st.file_uploader("📤 Upload a video file (e.g., MP4, AVI)", type=["mp4", "avi", "mov"])
-
-# Prediction button
-if uploaded_file is not None:
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as tmp_file:
-        tmp_file.write(uploaded_file.read())
-        tmp_file_path = tmp_file.name
-
-    st.info("⏳ Processing video... Please wait.")
-    
-    try:
-        # Initialize prediction class
-        predictor = Prediction()
-        
-        # Make prediction
-        prediction_result, gradcam_image, classification_details = predictor.predict(tmp_file_path)
-
-        # Display prediction
-        st.subheader("🧠 Prediction Result")
-        st.success(prediction_result)
-
-        # Display detailed classification info
-        if classification_details:
-            st.subheader("📊 Classification Details")
-            st.json(classification_details)
-
-        # Display Grad-CAM visualization
-        if gradcam_image is not None:
-            st.subheader("🔥 Grad-CAM Visualization (Middle Frame)")
-            st.image(gradcam_image, channels="BGR", caption="Grad-CAM Overlay")
-
-    except Exception as e:
-        st.error(f"❌ An error occurred during prediction: {e}")
-
-    finally:
-        # Clean up temp file
-        os.remove(tmp_file_path)
+if __name__ == "__main__":
+    demo.launch()