This project aims to extract information (title & description) from Boss Monster board game card's scans.
Be aware that this algorithm is not bulletproof and a human verification/correction is necessary.
Cargo takes care of all rusty dependencies.
However, leptess the rusty bindings for the lib Tesseract needs :
- development tesseract package
- development leptonica package
See leptess GitHub for more details.
cargo build --release
Get an overview of the supported arguments by running :
boss_monster_card_reader-cli --help
-i <path>add an input image path (multiple files can be added)-o <path>set the output file
To do
To get diagnostics features, crate features need to be enabled, see boss_monster_card_reader-core/Cargo.toml to get an exhaustive list.
From an acquisition point of view, card have to be placed on a white background with edge aligned with the scan borders.
All image format supported by the image crate can be used.
The algorithm virtually supports all languages as far as you have associated tesseract trained data, however, the language and path to data are still hard-coded in the boss_monster_card_reader-core::read_card function.
- JSON
Every format supported by the serde crate but, for the time, only JSON export have been added.