Decapod Roadmap

Table of Contents

Definitions

"Good Quality Source Material"

• segmented: distinguish lines and characters
• high enough contrast to detect characters
• skew should be corrected automatically or in the Decapod UI

Application Architecture

Hardware and image capture

• Two cameras
• Book flattening and camera alignment hardware
• Capture UX

or

• Flatbed scanner
• Folder full of images

or

• Single camera
• SD card
• Folder full of images
• or import UI for importing images from an SD card

Book Editing and Management Application

• Import images
• Reorder pages
• Delete pages
• Crop
• Rotation correction/skew
• Basic book metadata
• (Contrast controls)
• (Image annotation)

PDF Generation Pipeline

• Dewarping
• Binarization
• OCR
• PDF generation

Decapod Minimum Viable Application 1.0

• Import images
• Export:
  • Good quality image-based PDF
  • Searchable text in PDF
  • Text is selectable, screen readable, and extractable
  • Font matching

Supplementary Features

• Reorder pages
• Delete pages
• Basic book metadata (i.e. Dublin core fields)

Incremental Design/Development Roadmap

Decapod 0.5a

• For good quality source material, produce a PDF with:
  • Good image quality
  • Searchable text
  • Selectable, extractable text

Decapod 0.5

• New UI for importing a folder full of images (e.g. from a flat bed scanner)
• Polish export UI
• See Decapod 0.5 Planning

Decapod 0.6 (Full-featured Export to PDF)

Import / Export UI

• More export options: dimensions, DPI, colour depth, OCR text file output
• Report export status on the Exporter: make use of JSON status file
• TIFF output

Server

• Add support for new genpdf features
• Make use of genpdf JSON status file
• Report errors better (there are some generic errors)

Client and Server

• Make UI respond meaningfully to server error messages.
  • Requires some design.

Infrastructure

• Possible update to Ocropus 0.5
• Possible update to Ubuntu 12.04

ETA: 6 weeks (July 20th 2012)

• More detailed Decapod 0.6 planning

Decapod 0.7 (Single camera control)

• Single DSLR controlled within Decapod's user interface using gphoto
  • Detect attached camera.
  • Take a photo and transfer over USB to disk.

• See Decapod 0.7 Planning

Decapod 0.8

a. Stereo dewarping

• Physical rig for mounting two cameras sufficiently for stereo dewarping
• Calibration and capture UI

or

b. Structured light

• Physical rig for mounting a camera on top of a projector
• Calibration and capture UI

Next Steps

• Define criteria "good enough quality"
• Testing the limits of existing implementation
• Fix bugs in current PDF generation implementation

Things to Consider

• Metadata
  • basic book metadata
  • fancy dublin core, etc.
• Annotation
  • Text alternatives to images
  • translations
  • Scholarly annotations
  • Tagging

Things to consider

• New page management user interfaces
  • Deleting
  • Reordering
  • Cropping
  • Rotation
  • Undo Delete
  • Undo Crop, Rotation, Reordering
  • Redo Delete
  • Redo Crop, Rotation, Reordering
  • Ability to mark page boundaries within a single image of a two-page spread

• Book management
  • The ability to manage and work concurrently on more than one book at a time
  • Create, organize, delete
  • Edit information about the book (name, basic metadata, etc.)
  • Schedule books for export