Update app.py

app.py

@@ -1,15 +1,43 @@
-import gradio as gr
-from deepface import DeepFace
+import cv2
+import numpy as np
+
+def crop_face(image_path):
+    # Load the Haarcascade classifier for face detection
+    face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')
+
+    # Read the image
+    img = cv2.imread(image_path)
+    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+
+    # Detect faces in the image
+    faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5)
+
+    # If a face is detected, crop it
+    if len(faces) > 0:
+        (x, y, w, h) = faces[0]  # Take the first detected face
+        cropped_face = img[y:y+h, x:x+w]
+        return cropped_face
+    else:
+        raise ValueError("No face detected in the image.")
 
 def calculate_similarity(image1, image2):
-    # Analyze the two images for facial similarity
+    # Crop faces from both images
     try:
-        result = DeepFace.verify(image1, image2, enforce_detection=True)
-        similarity_percentage = result['distance'] * 100  # Higher distance means less similarity
-        return 100 - similarity_percentage  # Convert distance to similarity percentage
+        face1 = crop_face(image1)
+        face2 = crop_face(image2)
+
+        # Convert cropped images to RGB format for DeepFace
+        face1 = cv2.cvtColor(face1, cv2.COLOR_BGR2RGB)
+        face2 = cv2.cvtColor(face2, cv2.COLOR_BGR2RGB)
+
+        # Analyze the two cropped images for facial similarity
+        result = DeepFace.verify(face1, face2, model_name='VGG-Face', enforce_detection=False)
+        similarity_percentage = (1 - result['distance']) * 100  # Convert distance to similarity percentage
+        return similarity_percentage  
     except Exception as e:
         return str(e)
 
+
 # Create a Gradio interface
 iface = gr.Interface(
     fn=calculate_similarity,