Realize the gesture recognition of one complete sentance by the use of MHI and CNN. The accurancy can be 95%. 利用MHI來將影像轉換成靜態圖像,顏色深淺代表著時間先後順序 再利用CNN來辨識出每一個動作所代表的單詞 最後準確率可達95%, 並辨認出" I like my work "這類的句子