Update app.py

app.py

@@ -1,7 +1,5 @@
 from PIL import Image
 from transformers import BlipProcessor, BlipForConditionalGeneration
-import torch
-import cv2
 import numpy as np
 from deepface import DeepFace
 import gradio as gr
@@ -33,21 +31,15 @@ def analyze_image(image_pil):
     out = model.generate(**inputs)
     caption = processor.decode(out[0], skip_special_tokens=True)
 
-    # Face detection (Improved parameters + correct color conversion)
-    face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + "haarcascade_frontalface_default.xml")
-    gray = cv2.cvtColor(np.array(image_pil), cv2.COLOR_RGB2GRAY)  # Ensures color format is correct
-
-    # 🔍 More sensitive detection parameters
-    faces = face_cascade.detectMultiScale(
-        gray,
-        scaleFactor=1.05,  # increased sensitivity
-        minNeighbors=3,    # lower neighbor count = more detections
-        minSize=(30, 30)   # ensure even smaller faces are caught
-    )
+    # Face detection using DeepFace
+    try:
+        faces = DeepFace.extract_faces(img_path=image_pil, detector_backend="retinaface", enforce_detection=False)
+    except Exception:
+        faces = []
 
     face_infos = []
-    for (x, y, w, h) in faces:
-        face_crop = image_np[y:y+h, x:x+w]
+    for face_data in faces:
+        face_crop = face_data["face"]
         try:
             analysis = DeepFace.analyze(face_crop, actions=['age', 'gender', 'emotion'], enforce_detection=False)
             age = analysis[0]['age']
@@ -93,7 +85,7 @@ def analyze_image(image_pil):
     # Generate 15 sentences
     sentences = []
     sentences.append(f"According to the BLIP model, the scene can be described as: \"{caption}\".")
-    sentences.append(f"The image contains {num_faces} visible face(s) detected by OpenCV.")
+    sentences.append(f"The image contains {num_faces} visible face(s) detected using DeepFace (RetinaFace backend).")
 
     gender_desc = []
     if gender_counts["Man"] > 0: