Cropping investigation

copienator/paper_background.py

@@ -82,7 +82,7 @@ def _ruling(binary: np.ndarray, horizontal: bool) -> tuple[np.ndarray, int]:
 
 
 def _foreground(gray: np.ndarray, dpi: float, threshold: int,
-                ruling: np.ndarray) -> np.ndarray:
+                ruling: np.ndarray, *, diagnostics: dict | None = None) -> np.ndarray:
     block = max(15, int(dpi / 5) | 1)
     binary = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
                                    cv2.THRESH_BINARY_INV, block, threshold)
@@ -100,12 +100,14 @@ def _foreground(gray: np.ndarray, dpi: float, threshold: int,
         # locally thicker excess strokes inside the suppression mask.
         excess_ink = cv2.erode(excess_ink, np.ones((2, 2), np.uint8)) > 0
         mask &= ~excess_ink
+    if diagnostics is not None:
+        diagnostics['paper_suppressed'] = mask & (binary > 0)
     binary[mask] = 0
     return binary
 
 
 def _content_mask(binary: np.ndarray, dpi: float, nh: int = 0,
-                  nv: int = 0) -> tuple[np.ndarray, int]:
+                  nv: int = 0, *, diagnostics: dict | None = None) -> tuple[np.ndarray, int]:
     """Filter only tiny speckles and repeated, matching edge-hole components."""
     px = dpi / 25.4
     # Small closing reconnects strokes interrupted by paper-line suppression.
@@ -142,5 +144,10 @@ def _content_mask(binary: np.ndarray, dpi: float, nh: int = 0,
             holes.update(similar)
     if holes:
         keep[list(holes)] = False
+    if diagnostics is not None:
+        diagnostics['rejected_artifacts'] = ~keep[labels] & (binary > 0)
+        hole_labels = np.zeros(n, bool)
+        hole_labels[list(holes)] = True
+        diagnostics['repeated_holes'] = hole_labels[labels] & (binary > 0)
     # Bound original residual ink, not the expanded/grouped mask.
     return (keep[labels] & (binary > 0)).astype(np.uint8), len(holes)