p0,lo

app.py

@@ -39,13 +39,16 @@ app = Flask(__name__)
 CORS(app)
 
 def segment(image_pil: Image.Image, prompt: str):
-    # Convert PIL image to numpy array
-    image_np = np.array(image_pil)
+    # Convert PIL image to numpy array and normalize
+    image_np = np.array(image_pil).astype(np.float32) / 255.0  # Normalize to [0, 1]
+
+    # Convert numpy array to torch tensor
+    image_tensor = torch.tensor(image_np).permute(2, 0, 1).unsqueeze(0).to(device)  # Convert to CHW format
 
     # Run GroundingDINO to get boxes for the prompt
     boxes, _, _ = predict(
         model=grounder,
-        image=torch.tensor(image_np).to(device),  # Ensure image is on CPU
+        image=image_tensor,  # Pass normalized tensor
         caption=prompt,
         box_threshold=0.3,
         text_threshold=0.25,
@@ -56,7 +59,7 @@ def segment(image_pil: Image.Image, prompt: str):
 
     # 2) Largest box → mask via SAM
     box = boxes[np.argmax((boxes[:,2]-boxes[:,0])*(boxes[:,3]-boxes[:,1]))]
-    predictor.set_image(image_np)
+    predictor.set_image(np.array(image_pil))
     masks, _, _ = predictor.predict(box=box)
     mask = masks[0]  # boolean HxW