Update app.py

app.py

@@ -62,17 +62,17 @@ def create_user(email, password):
         print(f"User creation error: {str(e)}")
         return False, None
         
-# Update load_and_encode function to use the aligned face without normalization
+# Update load_and_encode function to return encodings for all detected faces
 @st_cache
 def load_and_encode(image_path):
     try:
-        aligned_face = detect_and_align_faces(image_path)
+        aligned_faces = detect_and_align_faces(image_path)
 
-        if aligned_face is not None:
-            encoding = face_recognition.face_encodings(aligned_face)
+        if aligned_faces:
+            encodings = face_recognition.face_encodings(aligned_faces)
 
-            if encoding:
-                return encoding
+            if encodings:
+                return encodings
             else:
                 return None
         else:
@@ -81,7 +81,7 @@ def load_and_encode(image_path):
         print(f"Error loading and encoding image: {str(e)}")
         return None
 
-# Function to detect and align faces in an image with preprocessing
+# Modify detect_and_align_faces function to detect and align multiple faces
 def detect_and_align_faces(image_path):
     image = face_recognition.load_image_file(image_path)
     
@@ -99,14 +99,14 @@ def detect_and_align_faces(image_path):
     if not faces:
         return None
 
-    # Use the first face found (you can modify this to handle multiple faces)
-    face = faces[0]
+    aligned_faces = []
+    for face in faces:
+        # Use dlib for face alignment
+        landmarks = shape_predictor(gray, face)
+        aligned_face = dlib.get_face_chip(resized_image, landmarks, size=256)  # Adjust the size as needed
+        aligned_faces.append(aligned_face)
     
-    # Use dlib for face alignment
-    landmarks = shape_predictor(gray, face)
-    aligned_face = dlib.get_face_chip(resized_image, landmarks, size=256)  # Adjust the size as needed
-    
-    return aligned_face
+    return aligned_faces
 
 # Add person to database
 @st_cache
@@ -137,24 +137,31 @@ def add_person(name, image_path, instagram_handle, email=None):
     except Exception as e:
         return f"Failed to add person: {str(e)}"
 
-# Recognize face from image
+# Update recognize_face function to handle multiple face encodings
 @st_cache
 def recognize_face(image_path):
     if not image_path:
         return "Please upload an image."
 
     try:
-        unknown_encoding = load_and_encode(image_path)
-        if not unknown_encoding:
+        unknown_encodings = load_and_encode(image_path)
+        if not unknown_encodings:
             return "No face found in the provided image."
 
         matches = []
-        for name, data in ref.get().items():
-            known_encoding = np.array(data["encoding"])
-            if face_recognition.compare_faces([known_encoding], unknown_encoding[0])[0]:
-                info = data["info"]
-                email = info.get("email", "Email not provided")
-                matches.append((name, info["instagram_handle"], email))
+        for unknown_encoding in unknown_encodings:
+            face_matches = []
+            for name, data in ref.get().items():
+                known_encoding = np.array(data["encoding"])
+                if face_recognition.compare_faces([known_encoding], unknown_encoding)[0]:
+                    info = data["info"]
+                    email = info.get("email", "Email not provided")
+                    face_matches.append((name, info["instagram_handle"], email))
+
+            if face_matches:
+                matches.extend(face_matches)
+            else:
+                matches.append(("Unknown", "Unknown", "Unknown"))
 
         if matches:
             results = []
@@ -169,38 +176,42 @@ def recognize_face(image_path):
     except Exception as e:
         return f"Failed to recognize face: {str(e)}"
         
-# Recognize face from image and return optimal or highest matching ID
+# Update recognize_face_optimal function to handle multiple face encodings
 @st_cache
 def recognize_face_optimal(image_path):
     if not image_path:
         return "Please upload an image."
 
     try:
-        unknown_encoding = load_and_encode(image_path)
-        if not unknown_encoding:
+        unknown_encodings = load_and_encode(image_path)
+        if not unknown_encodings:
             return "No face found in the provided image."
 
         matches = []
-        for name, data in ref.get().items():
-            known_encoding = np.array(data["encoding"])
-            similarity_score = face_recognition.face_distance([known_encoding], unknown_encoding[0])[0]
-            if similarity_score > 0.50:  # Only consider matches above 50.00% similarity
-                continue
-            matches.append((name, similarity_score))
+        for unknown_encoding in unknown_encodings:
+            face_matches = []
+            for name, data in ref.get().items():
+                known_encoding = np.array(data["encoding"])
+                similarity_score = face_recognition.face_distance([known_encoding], unknown_encoding)[0]
+                if similarity_score > 0.50:  # Only consider matches above 50.00% similarity
+                    continue
+                face_matches.append((name, similarity_score))
+
+            if face_matches:
+                best_match = min(face_matches, key=lambda x: x[1])
+                best_name, best_score = best_match
+                info = ref.child(best_name).child("info").get()
+                insta_handle = info["instagram_handle"]
+                insta_link = info["instagram_link"]
+                insta_link_html = f'<a href="{insta_link}" target="_blank"><font color="red">{insta_handle}</font></a>'
+                matches.append(f"Best match: {best_name} with a similarity score of {1 - best_score:.2%}. Insta handle: {insta_link_html}")
+            else:
+                matches.append("Face not found in the database.")
 
-        if matches:
-            best_match = min(matches, key=lambda x: x[1])
-            best_name, best_score = best_match
-            info = ref.child(best_name).child("info").get()
-            insta_handle = info["instagram_handle"]
-            insta_link = info["instagram_link"]
-            insta_link_html = f'<a href="{insta_link}" target="_blank"><font color="red">{insta_handle}</font></a>'
-            return f"Best match: {best_name} with a similarity score of {1 - best_score:.2%}. Insta handle: {insta_link_html}"
-        else:
-            return "Face not found in the database."
+        return "\n".join(matches)
     except Exception as e:
         return f"Failed to recognize face: {str(e)}"
-
+        
 # Delete person from database
 @st_cache
 def delete_person(name):
@@ -249,22 +260,12 @@ def recognize_face_ui():
         result = recognize_face(image_path)
         st.write(result, unsafe_allow_html=True)
 
-        if "It's a picture of" in result:
-            # Extract email from the result
-            email_start = result.find("Insta handle: ") + len("Insta handle: ")
-            email_end = result.find("</font></a>", email_start)
-            email = result[email_start:email_end]
-            st.write(f"Email: {email}")
-
 def recognize_face_optimal_ui():
     st.title("🔍 Recognize Face (Optimal)")
     image_path = st.file_uploader("Upload Image", help="Upload an image for optimal face recognition")
     if st.button("Recognize Face (Optimal)"):
         result = recognize_face_optimal(image_path)
-        if "not found" in result.lower():  # Check if "not found" is in the result message
-            st.error(result)
-        else:
-            st.write(result, unsafe_allow_html=True)
+        st.write(result, unsafe_allow_html=True)
         
 # Streamlit interface for deleting a person
 def delete_person_ui():