We are pleased to announce the release of the latest version of the ub-MOJI Dataset, version 25.09.
In this update, the following data have been newly added:
- Voiced sounds (e.g., ga, za, da, ba rows)
- Semi-voiced sounds (pa row)
- Long vowel mark (ー)
With this release, the dataset now covers a wider variety of finger spelling expressions. Future updates will continue to expand the coverage of voiced, semi-voiced, and long vowel signs, while also improving overall data quality.