手勢(shì)識(shí)別技術(shù)綜述(存儲(chǔ)版)
【正文】 2011【12】MISS. SHWETA K. YEWALE.ARTIFICIAL NEURAL NETWORK APPROACH FOR HAND GESTURE RECOGNITION.Shweta K. Yewale et al. / International Journal of Engineering Science and Technology (IJEST)。 所以無(wú)論是學(xué)生還是公司，只要堅(jiān)持對(duì)該系統(tǒng)的研究， 必然會(huì)獲得成功?！?0】到目前為止，大多數(shù)研究都集中在靜態(tài)手勢(shì)識(shí)別技術(shù)，而我們不僅要對(duì)手勢(shì)進(jìn)行跟蹤，還要進(jìn)行識(shí)別，其計(jì)算工作量很大且速度慢，不能用于實(shí)時(shí)識(shí)別系統(tǒng)。因?yàn)槭中斡袃煞N建模方式：基于三維的建模和基于圖像的建模。手勢(shì)識(shí)別要想取得比較高的識(shí)別率,仍有很長(zhǎng)的路要走?！?】 （2）用于手語(yǔ)識(shí)別。典型的人臉切爾諾夫模型擁有18個(gè)變量，而自1973年來(lái)經(jīng)過(guò)幾十年的發(fā)展，面部特征又得到了豐富，非對(duì)稱的切爾諾夫臉可顯示多達(dá)36維的臉部特征。某些具有手部整體運(yùn)動(dòng)軌跡的手勢(shì)可以是肢體動(dòng)作的一部分，比如揮手、打招呼等。由于其分析復(fù)雜，計(jì)算量大，速度慢，故而大多采用離散馬爾可夫模型?！?】大部分手勢(shì)識(shí)別應(yīng)用是將每個(gè)手勢(shì)作為一個(gè)整體，之后通過(guò)計(jì)算相似度來(lái)進(jìn)行模式匹配。該設(shè)備雖減輕了重量，提高了手部的靈活性，但是仍需較復(fù)雜的輸入輸出轉(zhuǎn)換設(shè)備，此外也會(huì)對(duì)手部動(dòng)作的自然性產(chǎn)生影響。手勢(shì)是任意的，手不同部位的方向、角度及彎曲程度等的不同信息可能會(huì)有實(shí)際意義上的天壤之別。Abstract:Gesture recognition is an interactive technology using mathematical arithmetic to the analysis,judge and assembly meaning that people want to convey which belongs to puter science and general, gesture recognition technology is not for simple gestures expressed by hands ,it can also aim to other body movement recognition, such as the head, arm and so on. But the gesture accounted for most of the analysis. In this paper, by describing the development process, tools used , objective and market of gesture recognition , we can sort out the ideas of the development of gesture recognition, and let readers have an overall understanding of gesture recognition. At the same time
醫(yī)療健康相關(guān)推薦
鄂ICP備17016276號(hào)-1