Update app.py

app.py

@@ -46,36 +46,24 @@ def parse_xml_points(xml_file):
     return np.array(points,dtype=np.float32).reshape(-1,2)
 
 # ---------------- Padding Helper ----------------
-def pad_to_match(img, target_h=None, target_w=None):
+def pad_to_size(img, target_h, target_w):
     h, w = img.shape[:2]
-    pad_top, pad_bottom, pad_left, pad_right = 0,0,0,0
-    if target_h is not None and h < target_h:
-        diff = target_h - h
-        pad_top = diff // 2
-        pad_bottom = diff - pad_top
-    if target_w is not None and w < target_w:
-        diff = target_w - w
-        pad_left = diff // 2
-        pad_right = diff - pad_left
-    canvas = np.ones((h+pad_top+pad_bottom, w+pad_left+pad_right,3), dtype=np.uint8)*255
-    canvas[pad_top:pad_top+h, pad_left:pad_left+w] = img
+    top_pad = 0
+    left_pad = 0
+    bottom_pad = target_h - h
+    right_pad = target_w - w
+    canvas = np.ones((target_h, target_w,3), dtype=np.uint8)*255
+    canvas[top_pad:top_pad+h, left_pad:left_pad+w] = img
     return canvas
 
-def merge_2x2_grid(images):
-    top_h = max(images[0].shape[0], images[1].shape[0])
-    bottom_h = max(images[2].shape[0], images[3].shape[0])
-    left_w = max(images[0].shape[1], images[2].shape[1])
-    right_w = max(images[1].shape[1], images[3].shape[1])
-
-    flat_pad  = pad_to_match(images[0], target_h=top_h, target_w=left_w)
-    match_pad = pad_to_match(images[1], target_h=top_h, target_w=right_w)
-    roi_pad   = pad_to_match(images[2], target_h=bottom_h, target_w=left_w)
-    xml_pad   = pad_to_match(images[3], target_h=bottom_h, target_w=right_w)
-
-    top_row = np.hstack([flat_pad, match_pad])
-    bottom_row = np.hstack([roi_pad, xml_pad])
-    grid = np.vstack([top_row, bottom_row])
-    return grid
+# ---------------- Resize feature-match to original reference size ----------------
+def match_img_to_reference(match_img, ref_h, ref_w):
+    h, w = match_img.shape[:2]
+    scale = min(ref_w/w, ref_h/h)
+    new_w, new_h = int(w*scale), int(h*scale)
+    resized = cv2.resize(match_img, (new_w,new_h))
+    padded = pad_to_size(resized, ref_h, ref_w)
+    return padded
 
 # ---------------- Main Function ----------------
 def homography_all_detectors(flat_file, persp_file, json_file, xml_file):
@@ -118,11 +106,25 @@ def homography_all_detectors(flat_file, persp_file, json_file, xml_file):
 
         # Convert to RGB
         flat_rgb = cv2.cvtColor(flat_img,cv2.COLOR_BGR2RGB)
-        match_rgb = cv2.cvtColor(match_img,cv2.COLOR_BGR2RGB)
+        persp_rgb = cv2.cvtColor(persp_img,cv2.COLOR_BGR2RGB)
         roi_rgb = cv2.cvtColor(persp_roi,cv2.COLOR_BGR2RGB)
         xml_rgb = cv2.cvtColor(xml_gt_img,cv2.COLOR_BGR2RGB)
 
-        combined_grid = merge_2x2_grid([flat_rgb, match_rgb, roi_rgb, xml_rgb])
+        # Resize feature-match image to match original flat/perspective
+        match_rgb = match_img_to_reference(cv2.cvtColor(match_img, cv2.COLOR_BGR2RGB), flat_rgb.shape[0], flat_rgb.shape[1])
+
+        # Determine max height and width for grid (all images now same)
+        max_h = max(flat_rgb.shape[0], match_rgb.shape[0], roi_rgb.shape[0], xml_rgb.shape[0])
+        max_w = max(flat_rgb.shape[1], match_rgb.shape[1], roi_rgb.shape[1], xml_rgb.shape[1])
+
+        flat_pad = pad_to_size(flat_rgb, max_h, max_w)
+        roi_pad = pad_to_size(roi_rgb, max_h, max_w)
+        xml_pad = pad_to_size(xml_rgb, max_h, max_w)
+
+        # Merge 2x2 grid
+        top = np.hstack([flat_pad, match_rgb])
+        bottom = np.hstack([roi_pad, xml_pad])
+        combined_grid = np.vstack([top, bottom])
 
         base_name = os.path.splitext(os.path.basename(persp_file))[0]
         file_name = f"{base_name}_{method.lower()}.png"
@@ -133,7 +135,6 @@ def homography_all_detectors(flat_file, persp_file, json_file, xml_file):
     while len(download_files)<5: download_files.append(None)
     return gallery_paths, download_files[0], download_files[1], download_files[2], download_files[3], download_files[4]
 
-# ---------------- Gradio UI ----------------
 iface = gr.Interface(
     fn=homography_all_detectors,
     inputs=[
@@ -151,7 +152,7 @@ iface = gr.Interface(
         gr.File(label="Download AKAZE Result")
     ],
     title="Homography ROI Projection with Feature Matching & XML GT",
-    description="Flat + Perspective images with mockup.json & XML. Original resolution maintained. Grid aligned with white padding."
+    description="Flat + Perspective images with mockup.json & XML. Feature-match aligned with original images using white padding."
 )
 
 iface.launch()