Under Community Review

Treat full-width and half-width characters as interchangeable for termbase recognition

When translating from Japanese to English, full-width numbers and characters should get hits in the termbase recognition for terms that were registered as a half-width charater/number (and vice versa). For example, if NG is registered in the termbase, then when the full-width ＮＧ is in the Japanese original text, the English equivalent for NG should show up in the term recognition window. 

Calvin Bulthuis

  • I have noticed the exact same problem. It would be great if termbase recognition could process half-width and full-width characters as the same character. The manual MultiTerm search function does not differentiate between the half-width and full-width versions of the Japanese term, but the termbase recognition does. Meaning that termbase recognition fails to show results for terms unless the original text uses the exact same character width version as the registered term.