Improving Hangul to English translation for optical character recognition (OCR)
An optical assessment method was proposed to help translation algorithms first assess whether the Korean text has been written as English syllables in Korean or in true Korean vocabulary before producing translated phrases.
Abstract
Real-time translation of languages using camera inputs occasionally results in awkward failures. As a proposed method of assisting such tools for Korean (Hangul) to English translation, an optical assessment method was proposed to help translation algorithms first assess whether the Korean text has been written as English syllables in Korean or in true Korean vocabulary before producing translated phrases. Although the current approach was not viable, future work will implement feedback regarding methods to meaningfully handle the optical data received.