Update app.py

app.py

@@ -1,55 +1,90 @@
 import gradio as gr
+import cv2
+import numpy as np
+import torch
 from transformers import AutoImageProcessor, SiglipForImageClassification
 from PIL import Image
-import torch
-import cv2
+import matplotlib.pyplot as plt
+import tempfile
 import os
-import uuid
 
-# Load model
+# ✅ Load model once
 model_name = "prithivMLmods/deepfake-detector-model-v1"
 processor = AutoImageProcessor.from_pretrained(model_name)
 model = SiglipForImageClassification.from_pretrained(model_name)
+model.eval()
 
-def analyze_video(video_path):
+# ✅ Load OpenCV Haar face detector
+face_detector = cv2.CascadeClassifier(cv2.data.haarcascades + "haarcascade_frontalface_default.xml")
+
+def analyze_deepfake(video_path):
     cap = cv2.VideoCapture(video_path)
-    result_labels = []
-    frame_skip = 10
-    count = 0
+    frame_preds = []
+    frame_count = 0
+    max_frames = 60
 
     while True:
         ret, frame = cap.read()
-        if not ret:
+        if not ret or frame_count >= max_frames:
             break
-        if count % frame_skip == 0:
-            try:
-                rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
-                pil = Image.fromarray(rgb)
-                inputs = processor(images=pil, return_tensors="pt")
-                with torch.no_grad():
-                    logits = model(**inputs).logits
-                    pred = torch.argmax(logits, dim=1).item()
-                    label = model.config.id2label[pred]
-                    result_labels.append(label)
-            except:
+
+        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
+        faces = face_detector.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5)
+
+        found = False
+        for (x, y, w, h) in faces:
+            face = frame[y:y+h, x:x+w]
+            if face.size == 0:
                 continue
-        count += 1
+
+            face_rgb = cv2.cvtColor(face, cv2.COLOR_BGR2RGB)
+            inputs = processor(images=Image.fromarray(face_rgb), return_tensors="pt")
+
+            with torch.no_grad():
+                logits = model(**inputs).logits
+                fake_prob = torch.softmax(logits, dim=-1)[0][1].item()
+
+            frame_preds.append(fake_prob)
+            found = True
+            break
+
+        if not found:
+            frame_preds.append(0.5)
+
+        frame_count += 1
 
     cap.release()
 
-    real = result_labels.count("REAL")
-    fake = result_labels.count("FAKE")
-    final = "REAL" if real > fake else "FAKE"
+    # Final Result
+    if frame_preds:
+        avg = np.mean(frame_preds)
+        verdict = "FAKE" if avg > 0.5 else "REAL"
+        result_text = f"✅ FINAL RESULT: **{verdict}** (confidence: {avg:.2f})"
+    else:
+        result_text = "❌ No faces detected. Please try another video."
+
+    # Plot histogram
+    fig, ax = plt.subplots(figsize=(6, 4))
+    ax.hist(frame_preds, bins=10, color="orange", edgecolor="black")
+    ax.set_title("Fake Confidence per Frame")
+    ax.set_xlabel("Confidence (0=Real, 1=Fake)")
+    ax.set_ylabel("Frame Count")
+    ax.grid(True)
+
+    # Save plot to temp file
+    plot_path = os.path.join(tempfile.gettempdir(), "plot.png")
+    plt.savefig(plot_path)
+    plt.close(fig)
 
-    return f"🟢 REAL frames: {real} | 🔴 FAKE frames: {fake} → Final verdict: **{final}**"
+    return result_text, plot_path
 
-# Gradio interface
+# ✅ Gradio Interface
 demo = gr.Interface(
-    fn=analyze_video,
-    inputs=gr.Video(label="Upload a video"),
-    outputs=gr.Markdown(label="Result"),
-    title="🎭 Deepfake Video Detector",
-    description="Upload a video (MP4). The model will analyze it and return whether it's REAL or FAKE based on detected face frames."
+    fn=analyze_deepfake,
+    inputs=gr.Video(label="📤 Upload a video (MP4 only)"),
+    outputs=[gr.Markdown(label="📊 Result"), gr.Image(type="filepath", label="📈 Confidence Histogram")],
+    title="🎭 Deepfake Video Detection App",
+    description="Upload a video. The model will detect faces and determine if it's REAL or FAKE using frame-level deepfake classification.",
 )
 
 demo.launch()