How to Import & Read
Explore the Webtoon reader, Side-by-Side comparison, Page Grid manager, Page Inspector, and Smart Re-slicing.
Preview Content
Please note that the details on this page are pre-content drafts. More detailed technical specifications, architectural breakdowns, and updated guides will be added soon.
1. Three Dedicated Chapter View Modes
Inside any chapter, XianScan provides 3 switchable view modes located on the top toolbar:
- Webtoon Continuous Reader: Infinite vertical scrolling with black background, optimized for continuous manhwa and webtoon strips. Includes a reading width selector (Compact
max-w-lg, Standardmax-w-2xl, Expandedmax-w-4xl) and a 1-click toggle between Translated Output and Original Raw. - Page Grid View: Visual thumbnail grid of all pages in the chapter. Displays live pipeline status badges (
Pending,Queued,Processing,Translated,Error), current pipeline step, and allows drag-and-drop page reordering. - Side-by-Side Compare View: Synchronized two-column layout placing the raw source page on the left and the translated canvas on the right for panel-by-panel verification.
2. Page Inspector Modal
Click on any page in Grid or Compare mode to open the Page Inspector Modal:
- 3-Layer Switching: Switch instantly between Output (Translated & Typeset), Cleaned (LaMa Neural Inpainting), and Original (Raw scan).
- OCR & Region Bounding Boxes: Inspect detected speech bubble polygons, OCR confidence scores, extracted source dialogue, and target translations.
- Single-Page Re-run: Re-run the translation or inpainting pipeline on an individual page without re-processing the entire chapter.
3. Smart Webtoon Re-slicing
Webtoons often suffer from arbitrary image splitting that cuts speech bubbles in half across slice boundaries.
Click Smart Re-slice in the chapter menu to:
- Recombine vertical strip fragments into a continuous image canvas.
- Detect non-text whitespace valleys and natural panel gutters.
- Automatically re-slice the chapter along clean panel boundaries before running speech bubble detection and OCR.