Update app.py

app.py

@@ -1,83 +1,18 @@
-import cv2
-import os
-import insightface
-import numpy as np
-from collections import deque
-
-# Initialize face detector and recognizer
-model = insightface.app.FaceAnalysis(allowed_modules=['detection', 'recognition'])
-model.prepare(ctx_id=0, det_size=(640, 640))  # Use ctx_id=-1 for CPU
-
-def normalize(v):
-    return v / np.linalg.norm(v)
-
-def cosine_similarity(a, b):
-    return np.dot(a, b)
-
-# Load known faces
-known_embs = []
-names = []
-for fname in os.listdir("images"):
-    if fname.lower().endswith(('.jpg', '.png')):
-        img = cv2.imread(os.path.join("images", fname))
-        faces = model.get(img)
-        if faces:
-            emb = normalize(faces[0].embedding)
-            known_embs.append(emb)
-            names.append(os.path.splitext(fname)[0])
-            print(f"Loaded {fname}")
-        else:
-            print(f"No face in {fname}")
-
-# Smoothing buffers for each face
-face_buffers = {}  # key: face_id, value: deque of embeddings
-
-# Start video
-cap = cv2.VideoCapture(0)
-face_id_counter = 0
-
-while True:
-    ret, frame = cap.read()
-    if not ret:
-        break
-
-    faces = model.get(frame)
-    current_buffers = {}
-
-    for i, face in enumerate(faces):
-        x1, y1, x2, y2 = face.bbox.astype(int)
-        emb = normalize(face.embedding)
-
-        # Use a temporary ID for each face based on bbox location
-        face_id = f"{x1}-{y1}-{x2}-{y2}"
-        if face_id not in face_buffers:
-            face_buffers[face_id] = deque(maxlen=5)
-
-        face_buffers[face_id].append(emb)
-        current_buffers[face_id] = face_buffers[face_id]  # mark as active this frame
-
-        # Smooth embedding
-        avg_emb = normalize(np.mean(face_buffers[face_id], axis=0))
-
-        # Find best match
-        sims = [cosine_similarity(avg_emb, known) for known in known_embs]
-        max_idx = np.argmax(sims)
-        name = "Unknown"
-        if sims[max_idx] > 0.5:
-            name = names[max_idx]
-
-        # Draw result
-        cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 200, 0), 2)
-        cv2.putText(frame, name, (x1, y1 - 10),
-                    cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 200, 0), 2)
-
-    # Remove stale buffers
-    face_buffers = {fid: buf for fid, buf in face_buffers.items() if fid in current_buffers}
-
-    # Show frame
-    cv2.imshow("InsightFace Multi-Face Recognition", frame)
-    if cv2.waitKey(1) == 27:  # ESC key
-        break
-
-cap.release()
-cv2.destroyAllWindows()
+import cv2
+import gradio as gr
+import numpy as np
+
+def detect_face(image):
+    # Load Haar cascade from OpenCV
+    face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')
+    img = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
+    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+
+    faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5)
+    for (x, y, w, h) in faces:
+        cv2.rectangle(img, (x, y), (x+w, y+h), (0, 255, 0), 2)
+
+    return cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+
+iface = gr.Interface(fn=detect_face, inputs=gr.Image(), outputs=gr.Image())
+iface.launch()