Using eye-tracking for real-time translation: a new approach to improving reading experience

In this paper, we consider the problem of comprehension difficulties caused by encountering new words in reading. Most of the existing ways to determine whether a user needs to translate a word are based on user-initiated clicks, but this is too cumbersome and has the possibility of mistakes. Therefore, this paper proposes a prediction method of word confusion based on eye tracking. This method uses the local perception of fixation method to find target words that the user may be confused about, then combines dynamic user eye-movement behavior data with static text information to build a Hidden Markov Model to predict whether the user is confused about these words, and finally selects a word with the highest confusion value and displays the words meaning in the annotation area. Experimental results show that the method achieves 87.4% precision in word confusion prediction, and the system's automatic annotation function improves users' reading efficiency and reading experience.