本研究結合現有的語音辨識引擎和語音合成引擎，再結合即時的影像追蹤功能，建置一套「電腦輔助英語訓練系統」，以增進國小學童英語聽、說能力。該系統的主要架構對於師生皆有輔助的功能，學童可根據自己的進度學習英語聽說能力，系統可發音讓學童學習，也可與學童模擬對話，並對學童的語音是否正確進行辨識，也可進行口說英語翻譯或中文翻譯。教師可透過系統視覺化的介面，進行教材內容之編輯，以及篩選學童的學習歷程資料，了解學童的學習進度與學習行為。實驗結果顯示，參與師生對系統的各項功能皆有正面的評價，本系統確實有效地增強學童在英語口說的能力，甚至在評鑑檢核學童的英語口說能力上，其正確性可作為英語學習護照認證的檢測評量工具。原本低成就的學童或對自己的英語口說較無自信的學童，大都能建立信心並通過英語學習護照認證。此外，對於專注力較差的學童也有激勵、督促的功效。教師有了系統的輔助，藉由學習歷程的紀錄分析，可輕易了解學童的學習狀況，有助於調整教學策略或個別補救教學，也減輕教師的負擔。本研究所提出之電腦輔助英語訓練系統之建置與應用模式，希望提供相關英語教學系統研發、教學應用實務及研究等之參考。

The study adopted the techniques of speech synthesis and speech recognition to improve English learning of elementary school students, assisted with computer vision lip recognition to observe and record the learning behavior of students. In order to develop an efficient learning system of English listening, speaking, reading, and translating, each module of learning system and learning behavior monitoring system was designed, implemented and experimented. From the verification of teaching and the findings of the experiment, the system is beneficial to improve students’ listening and speaking skills. In evaluating their speaking skills, it is more accurate than teachers. In assisting data analysis for teachers and supervising the learning progress and learning behavior of students, it also achieves the expected results.