Update app.py

app.py

@@ -40,49 +40,37 @@ transform = transforms.Compose([
 ])
 
 # -----------------------
-# Helper: basic fundus check
+# Helper: soft fundus check
 # -----------------------
 def looks_like_fundus(image):
     """
-    Heuristic to check if an image is likely a retinal fundus scan.
+    Very lightweight heuristic to guess if an image looks like a retinal fundus scan.
 
-    Assumptions:
-    - Fundus images usually have a bright circular central region (retina)
-      and a darker outer border (background).
-    - Non-fundus images (documents, labels, screens) tend to have
-      similar brightness across the whole frame.
+    This is NOT a medical-grade classifier – it is only used to show a warning
+    if the image is very unlikely to be a retina.
     """
-    # Grayscale + resize for consistency
     img = np.array(image.convert("L").resize((224, 224)))
 
-    # Central square (potential retina)
+    # Central square (potential retina) vs border
     center = img[40:184, 40:184]
 
-    # Border = everything outside the center
     border_mask = np.ones_like(img, dtype=bool)
     border_mask[40:184, 40:184] = False
     border_pixels = img[border_mask]
 
-    # Safety: if something weird happens, don't block it
+    # Safety fallback
     if border_pixels.size == 0:
         return True
 
     center_mean = center.mean()
     border_mean = border_pixels.mean()
 
-    # Ratios of dark/bright pixels
-    border_dark_ratio = np.mean(border_pixels < 40)     # dark background
-    center_bright_ratio = np.mean(center > 80)          # bright retina
+    border_dark_ratio = np.mean(border_pixels < 40)   # dark background
+    center_bright_ratio = np.mean(center > 80)        # bright retina
 
-    # Conditions (tunable):
-    # 1) Center clearly brighter than border
-    cond_contrast = center_mean - border_mean > 25
-
-    # 2) Most of the border is dark
-    cond_border_dark = border_dark_ratio > 0.5
-
-    # 3) A good portion of the center is bright
-    cond_center_bright = center_bright_ratio > 0.4
+    cond_contrast = center_mean - border_mean > 15
+    cond_border_dark = border_dark_ratio > 0.3
+    cond_center_bright = center_bright_ratio > 0.25
 
     return cond_contrast and cond_border_dark and cond_center_bright
 
@@ -91,14 +79,16 @@ def looks_like_fundus(image):
 # Predict and save
 # -----------------------
 def predict_retinopathy(image):
-    # 1. Block non-retina images BEFORE model / Grad-CAM
-    if not looks_like_fundus(image):
-        raise gr.Error(
-            "The uploaded image does not appear to be a retinal fundus scan. "
-            "Please upload a valid ophthalmic retinal image for Diabetic Retinopathy assessment."
+    # 1. Soft validation (no blocking – just warning text)
+    if looks_like_fundus(image):
+        warning = ""
+    else:
+        warning = (
+            "⚠️ The uploaded image may not be a retinal fundus scan. "
+            "This system is intended for use with ophthalmic retinal images only.\n\n"
         )
 
-    # 2. Normal pipeline for valid retinal images
+    # 2. Normal pipeline
     timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
     img = image.convert("RGB").resize((224, 224))
     img_tensor = transform(img).unsqueeze(0).to(device)
@@ -125,7 +115,9 @@ def predict_retinopathy(image):
     filename = f"{timestamp}_{label}_{confidence:.2f}.png"
     cam_pil.save(os.path.join(save_dir, filename))
 
-    return cam_pil, f"{label} (Confidence: {confidence:.2f})"
+    prediction_text = f"{label} (Confidence: {confidence:.2f})"
+
+    return cam_pil, warning + prediction_text
 
 
 # -----------------------
@@ -145,7 +137,8 @@ demo = gr.Interface(
     ),
     article=(
         "⚕️ **OpthaDetect** is an AI-powered ophthalmic decision-support tool. "
-        "It highlights retinal risk regions using Grad-CAM for better clinical interpretability."
+        "It highlights retinal risk regions using Grad-CAM for better clinical interpretability. "
+        "This tool does not replace clinical judgement and should be used alongside professional assessment."
     )
 )