feat: add table detection options and scan artifact removal

- Add TableDetectionSelector component for wired/wireless/region detection
- Add CV-based table line detector module (disabled due to poor performance)
- Add scan artifact removal preprocessing step (removes faint horizontal lines)
- Add PreprocessingConfig schema with remove_scan_artifacts option
- Update frontend PreprocessingSettings with scan artifact toggle
- Integrate table detection config into ProcessingPage
- Archive extract-table-cell-boxes proposal

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

backend/app/core/config.py

@@ -104,7 +104,15 @@ class Settings(BaseSettings):
     # Now using None to let PaddleX use its optimized defaults.
     layout_detection_threshold: Optional[float] = Field(default=None)  # None = use PaddleX default
     layout_nms_threshold: Optional[float] = Field(default=None)  # None = use PaddleX default
-    layout_merge_mode: Optional[str] = Field(default=None)  # None = use PaddleX default
+    # layout_merge_bboxes_mode options:
+    # - "large": Keep larger box when overlap (default)
+    # - "small": Keep smaller box when overlap
+    # - "union": Keep all boxes (preserve overlapping tables/images)
+    # Using "union" to prevent tables from being merged together
+    layout_merge_mode: Optional[str] = Field(
+        default="union",
+        description="How to handle overlapping detection boxes. 'union' preserves all detected regions."
+    )
     layout_unclip_ratio: Optional[float] = Field(default=None)  # None = use PaddleX default
 
     # Text Detection Parameters
@@ -161,13 +169,8 @@ class Settings(BaseSettings):
         description="Cell detection model for borderless tables. RT-DETR-L provides best accuracy."
     )
 
-    # Table Cell Boxes Extraction - supplement PPStructureV3 with direct SLANeXt calls
-    # When enabled, directly invokes SLANeXt models to extract cell bounding boxes
-    # which are not exposed by the PPStructureV3 high-level API
-    enable_table_cell_boxes_extraction: bool = Field(
-        default=True,
-        description="Enable direct SLANeXt model calls to extract table cell bounding boxes for accurate PDF layout."
-    )
+    # Note: Table cell boxes are now extracted from table_res_list returned by PPStructureV3
+    # No additional model calls needed - PPStructureV3 provides cell_box_list in table_res_list
 
     # Formula Recognition Model Configuration (Stage 4)
     # Available models: