国产日韩欧美一区二区三区三州_亚洲少妇熟女av_久久久久亚洲av国产精品_波多野结衣网站一区二区_亚洲欧美色片在线91_国产亚洲精品精品国产优播av_日本一区二区三区波多野结衣 _久久国产av不卡

?

基于加速度傳感器的WLALCS身份認(rèn)證及實(shí)現(xiàn)

2014-09-18 17:34高煥芝郭云鏑劉晴陳再良鄒北驥
關(guān)鍵詞:身份認(rèn)證云計(jì)算

高煥芝+郭云鏑+劉晴+陳再良+鄒北驥

收稿日期:20140326

基金項(xiàng)目:國(guó)家自然科學(xué)基金資助項(xiàng)目(61173122,61262032);湖南省自然科學(xué)基金資助項(xiàng)目(12JJ6059, 12JJ2038)

作者簡(jiǎn)介:高煥芝(1976-),女,河北唐山人,中南大學(xué)博士研究生

通訊聯(lián)系人,Email:csgrandeur@csu.edu.cn

摘要:針對(duì)當(dāng)前移動(dòng)設(shè)備身份認(rèn)證方法或易破解、或難實(shí)現(xiàn)、或成本高的問(wèn)題,提出一種新的基于手機(jī)加速度傳感器的人體感知身份認(rèn)證方法.該方法利用當(dāng)前手機(jī)普遍內(nèi)置的加速度傳感器采集人體運(yùn)動(dòng)數(shù)據(jù)(通常為動(dòng)態(tài)手勢(shì)),結(jié)合經(jīng)典匹配算法LCS,提出限制匹配窗口的近似判等最長(zhǎng)公共子序列算法,對(duì)采樣點(diǎn)序列限制區(qū)間匹配,針對(duì)浮點(diǎn)數(shù)據(jù)對(duì)采樣點(diǎn)距離近似判等,進(jìn)行數(shù)據(jù)匹配實(shí)現(xiàn)身份認(rèn)證,并基于云計(jì)算模型實(shí)現(xiàn)了手機(jī)身份認(rèn)證平臺(tái).較之已有的基于手勢(shì)身份認(rèn)證方法,有效降低了針對(duì)模仿動(dòng)作攻擊的接受錯(cuò)誤率,非攻擊認(rèn)證相等錯(cuò)誤率為2%,而模仿動(dòng)作攻擊的相等錯(cuò)誤率降低至5%.該系統(tǒng)具有易于在各類移動(dòng)設(shè)備系統(tǒng)部署,不需要額外的設(shè)備等優(yōu)勢(shì),且基于生物特征原理,特別加強(qiáng)了抵抗模仿動(dòng)作攻擊的健壯性,不易被破解.

關(guān)鍵詞:加速度傳感器;身份認(rèn)證;近似判等最長(zhǎng)公共子序列;云計(jì)算

中圖分類號(hào):TP391 文獻(xiàn)標(biāo)識(shí)碼:A

Accelerometer Based Authentication Method in WLALCS

GAO Huanzhi1, GUO Yundi1,2, LIU Qing1,2, CHEN Zailiang1,2, ZOU Beiji1,2

(1.School of Information Science and Engineering, Central South Univ,Changsha,Hunan410083, China;

2. Mobile Health Ministry of EducationChina Mobile Joint Laboratory, Changsha,Hunan410083, China )

Abstract:Aiming at problems in authentication methods of mobile devices, such as being easy to crack or difficult to implement or high costs, this paper presented a new mobile phone acceleration sensor authentication method on human perception. To use the current widespread acceleration sensor in mobile phones to capture human motion data (typically dynamic gesture), we proposed an authentication algorithm named Window Limited Approximate Longest Common Sequence (WLALCS) based on the classical matching algorithm LCS. And we implemented an authentication system on cloud computing model. Compared with the existing gesture and accelerometer based authentication methods, this method effectively reduced the equal error rate on imitate action attack. Nonimitate attack authentication EER (Equal Error Rate) is 2%, and imitate attack authentication EER is 5%. This system is easy to deploy on any smart phone systems and does not need any additional sensors. Based on the biological characteristics, we reinforced the robustness towards mimicry attack.

Key words:accelerometer; authentication; Window Limited Approximate Lonest Common Sequence(WLALCS) ; cloud computing

傳統(tǒng)的身份認(rèn)證方法有用戶名密碼、PIN、文本形式和九宮格等,此類方法輸入口令的時(shí)候很容易被竊取,從而仿冒真實(shí)用戶身份,簡(jiǎn)單的口令容易被破解,復(fù)雜的口令造成用戶記憶上的不便.因此,需要一種新的更為有效的身份認(rèn)證方式,要求簡(jiǎn)單易用、安全可靠.生物特征認(rèn)證技術(shù)的興起對(duì)傳統(tǒng)身份認(rèn)證方法起到了很好的補(bǔ)充和完善作用.

基于生物特征[1]的身份認(rèn)證,由于其安全、可靠和便利等特點(diǎn),越來(lái)越受到人們的重視,生物特征識(shí)別技術(shù)在過(guò)去的十幾年中取得了長(zhǎng)足的進(jìn)展,生物特征認(rèn)證方法主要有人臉認(rèn)證[2]、指紋認(rèn)證[3]、虹膜認(rèn)證[4]和手寫(xiě)簽名認(rèn)證[5]等.然而在手持設(shè)備身份認(rèn)證方面,以上認(rèn)證技術(shù)存在技術(shù)成本高、需要特定的硬件設(shè)施和易被仿冒等缺點(diǎn).

隨著2010年以來(lái)智能手機(jī)市場(chǎng)的爆發(fā)式增長(zhǎng),當(dāng)前的手機(jī)普遍帶有加速度傳感器(三軸加速度傳感器、陀螺儀等),同時(shí)結(jié)合手機(jī)本身能夠運(yùn)行軟件的特性,為低成本、高效率實(shí)現(xiàn)基于重力加速度傳感器的人體行為感知方法與身份認(rèn)證帶來(lái)了契機(jī).

近年來(lái)智能終端普遍裝備了越來(lái)越豐富的傳感器,出現(xiàn)了新的身份認(rèn)證方法.如利用三軸加速計(jì)進(jìn)行的步態(tài)身份認(rèn)證[6-8],利用三軸加速計(jì)把簡(jiǎn)單運(yùn)動(dòng)軌跡與密碼對(duì)應(yīng)的身份認(rèn)證[9-10]等.

每個(gè)人對(duì)持有的物體進(jìn)行空間移動(dòng)(通常為手持物體在空間中劃動(dòng)筆畫(huà),但不限于用手)時(shí),其動(dòng)作角度、力度和速度等都具有自己的個(gè)體特征,通過(guò)智能手機(jī)內(nèi)置加速度傳感器則可獲取用戶移動(dòng)手機(jī)時(shí)的這些空間動(dòng)作信息.針對(duì)用戶對(duì)手機(jī)完成的空間移動(dòng)數(shù)據(jù)的個(gè)體特征進(jìn)行分析,可實(shí)現(xiàn)身份認(rèn)證.這種認(rèn)證方法的好處是在數(shù)據(jù)采集過(guò)程中不依賴場(chǎng)景、光照、用戶體征完整性(視力、聽(tīng)力和語(yǔ)言等)以及額外的設(shè)備,實(shí)現(xiàn)成本低,方便易用,可用于需要身份認(rèn)證的場(chǎng)合,在即將進(jìn)入移動(dòng)互聯(lián)網(wǎng)的時(shí)代浪潮下,具有廣闊的應(yīng)用前景.

基于手勢(shì)的身份認(rèn)證方法,文獻(xiàn)[11]受到步態(tài)識(shí)別的啟發(fā),從生物特征的角度進(jìn)行了可行性分析,通過(guò)加速度傳感器采集手勢(shì)數(shù)據(jù),并通過(guò)監(jiān)督和非監(jiān)督降維進(jìn)行分類,通過(guò)數(shù)據(jù)分析,提出手勢(shì)運(yùn)動(dòng)可作為人體行為的生物特征,并可應(yīng)用于小群體的身份認(rèn)證的結(jié)論.文獻(xiàn)[12]結(jié)合文獻(xiàn)[13]中基于手勢(shì)識(shí)別的認(rèn)證方法在非模仿動(dòng)作攻擊場(chǎng)景下得到了良好的效果,對(duì)基于加速度傳感器采集手勢(shì)特征的身份認(rèn)證方法前景給出了積極的觀點(diǎn).文獻(xiàn)[13-15]描述了幾種基于手機(jī)三軸加速計(jì)進(jìn)行手勢(shì)身份認(rèn)證的方法,文獻(xiàn)[14]基于特定的手勢(shì)進(jìn)行認(rèn)證,不允許用戶自定義個(gè)性化的手勢(shì).文獻(xiàn)[15]中的方法需要較多的訓(xùn)練樣本才能達(dá)到較為理想的相等錯(cuò)誤率(Equal Error Rate,EER),文獻(xiàn)[13]中的工作更側(cè)重于人機(jī)交互的手勢(shì)識(shí)別精度,而認(rèn)證方面,在非攻擊認(rèn)證情況下取得了較好效果,但對(duì)于針對(duì)性模仿動(dòng)作攻擊,EER依然為10%,無(wú)法應(yīng)用于關(guān)鍵性信息的加密認(rèn)證.

移動(dòng)設(shè)備的身份認(rèn)證,要求具備易用性與實(shí)時(shí)性的特征,即要求僅需要較小的樣本,不受周邊環(huán)境影響,同時(shí)驗(yàn)證花費(fèi)時(shí)間短.需要大樣本訓(xùn)練的基于學(xué)習(xí)的方法(如HMM[16]等)、基于視覺(jué)(使用攝像頭)和語(yǔ)音等方法都會(huì)受到一定的限制.

本文基于動(dòng)態(tài)規(guī)劃原理、最長(zhǎng)公共子序列(Longest Common Sequence, LCS)[17]算法,設(shè)計(jì)了一個(gè)限制匹配窗口的近似判等最長(zhǎng)公共子序列(Window Limited Approximate Longest Common Sequence, WLALCS)算法,并以動(dòng)態(tài)時(shí)間規(guī)整(Dynamic Time Warping, DTW)[18]作為輔助判斷條件,在Lumia1520和PC服務(wù)器上實(shí)現(xiàn)了基于云計(jì)算的手機(jī)身份認(rèn)證原型系統(tǒng).該系統(tǒng)整個(gè)認(rèn)證過(guò)程在50 ms左右,系統(tǒng)不受光照和噪聲等環(huán)境影響,通過(guò)20名20~30歲的實(shí)驗(yàn)者一周時(shí)間進(jìn)行手勢(shì)動(dòng)作攻擊實(shí)驗(yàn)結(jié)果,該系統(tǒng)有效將模仿動(dòng)作攻擊的認(rèn)證過(guò)程相等錯(cuò)誤率(EER)降低至5%,有效地改善了基于三軸加速計(jì)的手勢(shì)身份認(rèn)證方法在模仿動(dòng)作攻擊時(shí)的不健壯性.

1WLALCS認(rèn)證算法

本文針對(duì)利用加速度傳感器提取動(dòng)作特征進(jìn)行身份認(rèn)證問(wèn)題,基于LCS設(shè)計(jì)了WLALCS算法,同時(shí)以DTW算法作為針對(duì)模板與驗(yàn)證數(shù)據(jù)序列累積距離輔助判斷條件.

1.1數(shù)據(jù)獲取及處理

利用Lumia搭載的WP8系統(tǒng)內(nèi)置三軸加速度傳感器獲取運(yùn)動(dòng)數(shù)據(jù),結(jié)合軟件設(shè)置,用戶將按鈕按下為動(dòng)作起點(diǎn),系統(tǒng)開(kāi)始記錄用戶動(dòng)作的加速度數(shù)據(jù);按鈕松開(kāi)為手勢(shì)終點(diǎn),系統(tǒng)停止記錄.

采樣頻率為60 Hz,獲取數(shù)據(jù)為時(shí)間序列的x,y和z三軸加速度g(g為重力加速度,取9.8 m/s2)的浮點(diǎn)數(shù),并對(duì)數(shù)據(jù)平滑處理以消除沖擊噪聲.

1.2WLALCS算法思想

對(duì)于時(shí)間序列X,設(shè)X.length為X的采樣點(diǎn)個(gè)數(shù),X[i]為時(shí)間順序第i個(gè)采樣點(diǎn).設(shè)待匹配序列T為模板序列,S為樣本序列.

1.2.1數(shù)據(jù)篩選

由于人完成特定動(dòng)作的時(shí)間有一定的不穩(wěn)定性,而人類的反應(yīng)時(shí)間一般在0.3 s,針對(duì)采樣頻率, 0.3×60 = 18,實(shí)驗(yàn)時(shí)取允許采樣點(diǎn)個(gè)數(shù)誤差為20,即對(duì)于長(zhǎng)度與模板相差超過(guò)20的數(shù)據(jù)將直接拒絕.

1.2.2最長(zhǎng)公共子序列(LCS)

最長(zhǎng)公共子序列算法是基于動(dòng)態(tài)規(guī)劃的經(jīng)典線性序列匹配算法,子序列為線性序列順序連續(xù)或非連續(xù)子集,最長(zhǎng)公共子序列指兩個(gè)或多個(gè)線性序列的最長(zhǎng)非連續(xù)公共部分,可有效表示兩個(gè)線性序列的編輯距離和相似度[17].

算法描述如下[19]:

兩個(gè)序列T和S的最長(zhǎng)公共子序列為L(zhǎng)CS(T,S),LCSdp(i,j)表示由序列T的T[1]~T[i]組成的子序列與序列S的S[1]~S[j]組成的子序列的最長(zhǎng)公共部分,即該動(dòng)態(tài)規(guī)劃算法的子問(wèn)題,如式(1)所示,LCSdp(T.length,S.length)即為L(zhǎng)CS(T,S)的結(jié)果.

LCST,S=LCSdpT.length,S.length.(1)

由規(guī)則EQUAL(T[i],S[i])判定T[i]與S[i]相等關(guān)系.當(dāng)EQUAL(T[i],S[i])分別為TRUE和FALSE時(shí),LCSdp采取不同的遞推規(guī)則.EQUAL(T[i],S[i])為TRUE時(shí),規(guī)則如式(2)所示,為FALSE時(shí),規(guī)則為式(3)所示.

LCSdpi,j=maxLCSdpi-1,j-1+1

LCSdpi-1,j

LCSdpi,j-1, (2)

LCSdpi,j=maxLCSdpi-1,j

LCSdpi,j-1.(3)

1.2.3限制匹配窗口

由于人有一定的反應(yīng)時(shí)間,故模板和測(cè)試數(shù)據(jù)序列的動(dòng)作起點(diǎn)往往是沒(méi)有完全對(duì)齊的,在做LCS時(shí)枚舉匹配起點(diǎn),并限制匹配窗口大小,取最長(zhǎng)匹配結(jié)果為最終匹配結(jié)果,即Window limited.取T和S采樣點(diǎn)較少的一個(gè)的采樣點(diǎn)個(gè)數(shù)為匹配窗口大小,這里假設(shè)S.length < T.length,枚舉匹配起點(diǎn)start由1至T.length-S.length,設(shè)每個(gè)start對(duì)應(yīng)的T[start]~T[start+S.length-1]的采樣點(diǎn)組成的序列為P,則WLALCS(T,S)為:

WLALCST,S=max[LCS(P,S)]. (4)

1.2.4近似判等

對(duì)于式(2),由于獲取的數(shù)據(jù)序列為浮點(diǎn)數(shù),對(duì)于相近的采樣點(diǎn)很難完全相等,故可規(guī)定兩點(diǎn)距離小于一定閾值時(shí)判定兩點(diǎn)相等,即近似判等.

設(shè)定ratio_errdis判等閾值,定義S[i]與T[i]的距離為三軸坐標(biāo)差的絕對(duì)值與T所有采樣點(diǎn)對(duì)應(yīng)坐標(biāo)最大值和最小值的差的比值的和,即

DistanceTi,Si=

∑x,y,zpTi.p-Si.pmaxT.p-minT.p.(5)

則當(dāng)Distance(T[i],S[i])

1.2.5序列相似度

定義序列相似度為WLALCS(T,S)的匹配長(zhǎng)度與模版T的長(zhǎng)度的比值.即序列相似度matched_proportion為:

matched_proportioin=WLALCST,ST.length. (6)

1.2.6算法復(fù)雜度

由于采樣點(diǎn)個(gè)數(shù)與模板相差20以上的樣本將被直接排除,故枚舉起點(diǎn)的復(fù)雜度可視為常數(shù),WLALCS的時(shí)間和空間復(fù)雜度都為O(T.length·S.length).

1.3動(dòng)態(tài)時(shí)間規(guī)整

動(dòng)態(tài)時(shí)間規(guī)整(DTW)是基于動(dòng)態(tài)規(guī)劃的對(duì)線性序列進(jìn)行模式匹配的經(jīng)典算法,文獻(xiàn)[13]在手勢(shì)識(shí)別方法上對(duì)該算法有了詳細(xì)的描述,本研究使用此方法作為輔助判斷方法,閾值參數(shù)會(huì)相對(duì)放寬.算法簡(jiǎn)單描述如下:

DTW(T,S)為序列T與S的DTW距離,DTWdp(i,j)表示由序列T的T[1]~T[i]組成的子序列與序列S的S[1]~S[j]組成的子序列的最優(yōu)DTW距離,DTWdp(T.length,S.length)即為DTW(T,S)的結(jié)果,如式(7)所示.最優(yōu)子問(wèn)題遞推規(guī)則如式(8)所示.

DTWT,S=DTWdpT.length,S.length. (7)

DTWdpi,j=minDTWdpi-1,j-1

DTWdpi-1,j

DTWdpi,j-1+

DistanceTi,Sj. (8)

DTW的時(shí)間和空間復(fù)雜度都為O(T.length·S.length).

2云計(jì)算認(rèn)證平臺(tái)

目前流行的移動(dòng)設(shè)備操作系統(tǒng)復(fù)雜多樣,主流的有Android, IOS, Windows Phone等,但即使相同的Android系統(tǒng),也因?yàn)榘姹静煌?,給APP開(kāi)發(fā)者的軟件適配帶來(lái)很多麻煩.同時(shí),現(xiàn)有設(shè)備的CPU浮點(diǎn)運(yùn)算能力與3~5年前的PC機(jī)相仿,雖然運(yùn)行論文所提算法沒(méi)有問(wèn)題,但對(duì)于未來(lái)提高認(rèn)證精度而可能會(huì)采用的更為復(fù)雜的算法,移動(dòng)設(shè)備的硬件壓力將越來(lái)越大.目前移動(dòng)互聯(lián)網(wǎng)已進(jìn)入3G時(shí)代,而4G技術(shù)也正在興起.本文所采集的每組手勢(shì)數(shù)據(jù)僅為1~5 kB,足夠在普通應(yīng)用場(chǎng)所進(jìn)行快速數(shù)據(jù)交換.

云計(jì)算是一種新興的計(jì)算模型,對(duì)用戶透明,用戶無(wú)需了解云計(jì)算的具體機(jī)制即可獲得需要的服務(wù)[20].基于云計(jì)算認(rèn)證平臺(tái)的實(shí)現(xiàn),使得客戶端的開(kāi)發(fā)只需要考慮數(shù)據(jù)的采集和發(fā)送,整個(gè)認(rèn)證過(guò)程在云端進(jìn)行,極大地提高了開(kāi)發(fā)效率和身份認(rèn)證算法的運(yùn)行效率,降低APP開(kāi)發(fā)的適配壓力,解放了移動(dòng)設(shè)備CPU和存儲(chǔ)設(shè)備,對(duì)未來(lái)利用云端的手勢(shì)數(shù)據(jù)進(jìn)行大數(shù)據(jù)分析也帶來(lái)了便利.

2.1總體設(shè)計(jì)

客戶端為移動(dòng)設(shè)備,服務(wù)端為計(jì)算機(jī)服務(wù)器.客戶端將采集的手勢(shì)數(shù)據(jù)發(fā)送至服務(wù)器,在服務(wù)器進(jìn)行匹配認(rèn)證,并將認(rèn)證數(shù)據(jù)存儲(chǔ)在數(shù)據(jù)服務(wù)器,以供未來(lái)研究分析.并返回結(jié)果給客戶端.數(shù)據(jù)傳輸遵循TCP/IP協(xié)議,客戶端通過(guò)無(wú)線網(wǎng)發(fā)送數(shù)據(jù),系統(tǒng)總體設(shè)計(jì)如圖2所示.系統(tǒng)自頂向下分為數(shù)據(jù)處理云端、數(shù)據(jù)交換接口、客戶端、用戶,系統(tǒng)結(jié)構(gòu)如圖3所示.

2.2實(shí)現(xiàn)

2.2.1數(shù)據(jù)處理云端

數(shù)據(jù)處理云端,實(shí)驗(yàn)采用普通酷睿雙核,4 GB內(nèi)存PC機(jī),基于WPF開(kāi)發(fā)的集數(shù)據(jù)接收發(fā)送、存儲(chǔ)、管理、匹配認(rèn)證、圖形化分析于一體的服務(wù)器軟件,能夠?qū)崟r(shí)偵聽(tīng)用戶數(shù)據(jù),存儲(chǔ)歷史認(rèn)證數(shù)據(jù),對(duì)用戶認(rèn)證信息進(jìn)行快速反饋.

2.2.2數(shù)據(jù)交換接口

客戶端與服務(wù)器的數(shù)據(jù)交換使用TCP/IP協(xié)議,通過(guò)客戶端與服務(wù)器建立TCP鏈接.客戶端采集的手勢(shì)數(shù)據(jù)以文件流形式通過(guò)TCP鏈接發(fā)送至服務(wù)器,數(shù)據(jù)格式為:

1)指令行,標(biāo)定數(shù)據(jù)文件處理方式,以@開(kāi)頭.

2)用戶信息,標(biāo)定發(fā)送數(shù)據(jù)的用戶.

3)手勢(shì)數(shù)據(jù).

2.2.3客戶端

客戶端通過(guò)調(diào)用傳感器的API獲取用戶手勢(shì)數(shù)據(jù),在手勢(shì)結(jié)束后對(duì)數(shù)據(jù)打包進(jìn)行發(fā)送.

3實(shí)驗(yàn)結(jié)果與分析

本文針對(duì)認(rèn)證系統(tǒng)對(duì)模仿動(dòng)作攻擊的健壯性,設(shè)計(jì)不同組別的手勢(shì),并令參與實(shí)驗(yàn)人員了解手勢(shì)的執(zhí)行方式和執(zhí)行時(shí)間,并且所有實(shí)驗(yàn)者完全在視覺(jué)暴露的環(huán)境下執(zhí)行手勢(shì),互相作為模仿動(dòng)作攻擊對(duì)象.

3.1實(shí)驗(yàn)結(jié)果

通過(guò)對(duì)序列相似度認(rèn)證閾值threshold的研究,對(duì)于模仿動(dòng)作攻擊數(shù)據(jù),受試者工作特征(receiver operating characteristic,ROC)曲線如圖4所示.

圖4中,F(xiàn)AR為認(rèn)假率(False Accept Rate),F(xiàn)RR為拒真率(False Reject Rate),對(duì)于模仿動(dòng)作攻擊認(rèn)證,EER達(dá)到5%,優(yōu)于文獻(xiàn)[7]和[8]中的10%.對(duì)于非模仿動(dòng)作攻擊認(rèn)證,所提方法的EER也達(dá)到了2%,優(yōu)于文獻(xiàn)[7]和[8]中的3%.認(rèn)證效果對(duì)比如表1所示.

3.2實(shí)驗(yàn)分析

由圖4可知,隨著threshold值的增加,用戶的認(rèn)證難度加大,同時(shí)對(duì)模仿動(dòng)作攻擊的防御能力也隨之增強(qiáng).而在模仿動(dòng)作攻擊成功率降低至較低水平時(shí),用戶認(rèn)證的成功率依然比較可觀.

不同的人做手勢(shì)的時(shí)候,由于個(gè)人習(xí)慣、肌肉強(qiáng)度和骨骼結(jié)構(gòu)不同等方面的影響,動(dòng)作不同階段特別是變換速度與方向的時(shí)候會(huì)有較大的差別,大多數(shù)情況下完成一個(gè)動(dòng)作的時(shí)間乃至一個(gè)動(dòng)作的不同階段的時(shí)間會(huì)有不同,僅從采樣點(diǎn)累積距離角度出發(fā),不能很好地反映這方面的差異.論文所提算法采用了限制窗口的近似最長(zhǎng)公共子序列方法,有效地反映了模板與測(cè)試數(shù)據(jù)之間局部運(yùn)動(dòng)速度特點(diǎn)的差異和整體動(dòng)作的采樣點(diǎn)匹配比例,在模仿動(dòng)作欺騙攻擊的防御上起到了更好的效果.

4總結(jié)

針對(duì)基于手機(jī)加速度傳感器進(jìn)行身份認(rèn)證方法中,對(duì)于模擬動(dòng)作攻擊的健壯性不強(qiáng)的問(wèn)題,進(jìn)行了算法研究,提出WLALCS算法在保持正常認(rèn)證環(huán)境下的認(rèn)證精度的前提下,有效提高了模擬動(dòng)作攻擊的抵抗力.

通過(guò)模擬動(dòng)作攻擊實(shí)驗(yàn),建立了近2 000個(gè)手勢(shì)的模仿動(dòng)作攻擊手勢(shì)數(shù)據(jù)庫(kù),通過(guò)實(shí)驗(yàn),所提算法WLALCS的EER達(dá)到了2%,模仿動(dòng)作攻擊EER達(dá)到了5%,對(duì)于模仿攻擊具有了更好的防御性,提高了認(rèn)證系統(tǒng)的安全性.

研究基于NOKIA Lumia1520,個(gè)人計(jì)算機(jī)實(shí)現(xiàn)了云計(jì)算模型認(rèn)證系統(tǒng).WLALCS算法結(jié)合基于云計(jì)算模型的認(rèn)證系統(tǒng),有效減輕了不同系統(tǒng)移動(dòng)終端開(kāi)發(fā)的適配壓力,解放了移動(dòng)終端的CPU和存儲(chǔ)壓力,認(rèn)證算法的運(yùn)行效率不再受限于移動(dòng)終端CPU配置的高低,為未來(lái)實(shí)現(xiàn)更為復(fù)雜精細(xì)的認(rèn)證系統(tǒng)提供了條件.認(rèn)證歷史數(shù)據(jù)存儲(chǔ)在云端,也使針對(duì)動(dòng)作特征的大數(shù)據(jù)分析成為可能.

參考文獻(xiàn)

[1]PANKANTI S, BOLLE R,JAIN A.Biometrics: the future of identification[J]. Computer, 2000, 33(2): 46-49.

[2]WALLACE R,MCLAREN M,MCCOOL C,et al. Crosspollination of normalization techniques from speaker to face authentication using gaussian mixture models[J]. Information Forensics and Security, IEEE Transactions on, 2012, 7(2): 553-562.

[3]NANNI L, BRAHNAM S, LUMINI A. Biohashing applied to orientationbased minutia descriptor for secure fingerprint authentication system[J]. Electronics letters, 2011, 47(15): 851-853.

[4]BANERJEE S, MUKHOPADHYAY S, RONDONI L. Multiimage encryption based on synchronization of chaotic lasers and iris authentication[J]. Optics and Lasers in Engineering, 2012, 50(7): 950-957.

[5]SCHEIDAT T, KMMEL K, VIELHAUER C. Short term template aging effects on biometric dynamic handwriting authentication performance[C]//Communications and Multimedia Security. Berlin:Springer Berlin Heidelberg, 2012: 107-116.

[6]GAFUROV D, HELKALA K, SNDROL T. Biometric gait authen tication using accelerometer sensor[J]. Journal of computers, 2006, 1(7): 51-59.

[7]TAMVIRUZZAMAN M, AHAMED S I, HASAN C S,et al. ePet: when cellular phone learns to recognize its owner[C]//Proceedings of the 2nd ACM Workshop on Assurable and Usable Security Configuration. ACM, 2009: 13-18.

[8]DERAWI M,NICKEL C,BOURS P,et al.Unobtrusive userauthentication on mobile phones using biometric gait recognition[C]//Intelligent Information Hiding and Multimedia Signal Processing (IIHMSP), 2010 Sixth International Conference on. Piscataway, N J: IEEE, 2010: 306-311.

[9]PATEL S N,PIERCE J,ABOWD G.A gesturebased authentication scheme for untrusted public terminals[C]//Proceedings of the 17th Annual ACMS Ymposium on User Interface Software and Technology. New York: ACM Press ,2004:157-160.

[10]CHONG M,MARSDEN G.Exploring the use of discrete gestures for authentication[M]. Berlin: Springer Berlin Heidelberg, 2009: 205-213.

[11]FARELLA E,OMODHRAIN S,BENINI L, et al. Gesture signature for ambient intelligence applications: a feasibility study[M]. Berlin: Springer Berlin Heidelberg, 2006: 288-304.

[12]LIU J, ZHONG L, WICKRAMASURIYA J, et al. User evaluation of lightweight user authentication with a single triaxis accelerometer[C]//Proceedings of the 11th International Conference on HumanComputer Interaction with Mobile Devices and Services. NewYork: ACM, 2009: 15.

[13]LIU J, ZHONG L, WICKRAMASURIYA J, et al. uWave: accelerometerbased personalized gesture recognition and its applications[J]. Pervasive and Mobile Computing, 2009, 5(6): 657-675.

[14]OKUMURA F, KUBOTA A, HATORI Y, et al. A study on biometric authentication based on arm sweep action with acceleration sensor[C]//Intelligent Signal Processing and Communications.ISPACS'06,International Symposium on. Piscataway, N J: IEEE, 2006: 219-222.

[15]MATSUO K, OKUMURA F, HASHIMOTO M, et al. Arm swing identification method with template update for long term stability[M]. Berlin: Springer Berlin Heidelberg, 2007: 211-221.

[16]RABINER L, JUANG B. An introduction to hidden Markov models[J]. ASSP Magazine, IEEE, 1986, 3(1): 4-16.

[17]HUNT J,SZYMANSKI T.A fast algorithm for computing longest common subsequences[J]. Communications of the ACM, 1977, 20(5): 350-353.

[18]MYERS C,HABINER L.A Comparative study of several dynamic timewarping algorithms for connected word[J]. Bell System Technical Journal, 1981,60(7):1389-1409.

[19]CORMEN T,LEISERSON C,RIVEST R,et al.Introduction to algorithms[M]. 3rd ed. Cambridge: MIT Press, 2001:390-397.

[20]ARMBRUST M, FOX A, GRIFFITH R, et al. A view of cloud computing[J]. Communications of the ACM, 2010, 53(4): 50-58.

[10]CHONG M,MARSDEN G.Exploring the use of discrete gestures for authentication[M]. Berlin: Springer Berlin Heidelberg, 2009: 205-213.

[11]FARELLA E,OMODHRAIN S,BENINI L, et al. Gesture signature for ambient intelligence applications: a feasibility study[M]. Berlin: Springer Berlin Heidelberg, 2006: 288-304.

[12]LIU J, ZHONG L, WICKRAMASURIYA J, et al. User evaluation of lightweight user authentication with a single triaxis accelerometer[C]//Proceedings of the 11th International Conference on HumanComputer Interaction with Mobile Devices and Services. NewYork: ACM, 2009: 15.

[13]LIU J, ZHONG L, WICKRAMASURIYA J, et al. uWave: accelerometerbased personalized gesture recognition and its applications[J]. Pervasive and Mobile Computing, 2009, 5(6): 657-675.

[14]OKUMURA F, KUBOTA A, HATORI Y, et al. A study on biometric authentication based on arm sweep action with acceleration sensor[C]//Intelligent Signal Processing and Communications.ISPACS'06,International Symposium on. Piscataway, N J: IEEE, 2006: 219-222.

[15]MATSUO K, OKUMURA F, HASHIMOTO M, et al. Arm swing identification method with template update for long term stability[M]. Berlin: Springer Berlin Heidelberg, 2007: 211-221.

[16]RABINER L, JUANG B. An introduction to hidden Markov models[J]. ASSP Magazine, IEEE, 1986, 3(1): 4-16.

[17]HUNT J,SZYMANSKI T.A fast algorithm for computing longest common subsequences[J]. Communications of the ACM, 1977, 20(5): 350-353.

[18]MYERS C,HABINER L.A Comparative study of several dynamic timewarping algorithms for connected word[J]. Bell System Technical Journal, 1981,60(7):1389-1409.

[19]CORMEN T,LEISERSON C,RIVEST R,et al.Introduction to algorithms[M]. 3rd ed. Cambridge: MIT Press, 2001:390-397.

[20]ARMBRUST M, FOX A, GRIFFITH R, et al. A view of cloud computing[J]. Communications of the ACM, 2010, 53(4): 50-58.

[10]CHONG M,MARSDEN G.Exploring the use of discrete gestures for authentication[M]. Berlin: Springer Berlin Heidelberg, 2009: 205-213.

[11]FARELLA E,OMODHRAIN S,BENINI L, et al. Gesture signature for ambient intelligence applications: a feasibility study[M]. Berlin: Springer Berlin Heidelberg, 2006: 288-304.

[12]LIU J, ZHONG L, WICKRAMASURIYA J, et al. User evaluation of lightweight user authentication with a single triaxis accelerometer[C]//Proceedings of the 11th International Conference on HumanComputer Interaction with Mobile Devices and Services. NewYork: ACM, 2009: 15.

[13]LIU J, ZHONG L, WICKRAMASURIYA J, et al. uWave: accelerometerbased personalized gesture recognition and its applications[J]. Pervasive and Mobile Computing, 2009, 5(6): 657-675.

[14]OKUMURA F, KUBOTA A, HATORI Y, et al. A study on biometric authentication based on arm sweep action with acceleration sensor[C]//Intelligent Signal Processing and Communications.ISPACS'06,International Symposium on. Piscataway, N J: IEEE, 2006: 219-222.

[15]MATSUO K, OKUMURA F, HASHIMOTO M, et al. Arm swing identification method with template update for long term stability[M]. Berlin: Springer Berlin Heidelberg, 2007: 211-221.

[16]RABINER L, JUANG B. An introduction to hidden Markov models[J]. ASSP Magazine, IEEE, 1986, 3(1): 4-16.

[17]HUNT J,SZYMANSKI T.A fast algorithm for computing longest common subsequences[J]. Communications of the ACM, 1977, 20(5): 350-353.

[18]MYERS C,HABINER L.A Comparative study of several dynamic timewarping algorithms for connected word[J]. Bell System Technical Journal, 1981,60(7):1389-1409.

[19]CORMEN T,LEISERSON C,RIVEST R,et al.Introduction to algorithms[M]. 3rd ed. Cambridge: MIT Press, 2001:390-397.

[20]ARMBRUST M, FOX A, GRIFFITH R, et al. A view of cloud computing[J]. Communications of the ACM, 2010, 53(4): 50-58.

猜你喜歡
身份認(rèn)證云計(jì)算
云電子身份管理與認(rèn)證系統(tǒng)中的關(guān)鍵技術(shù)優(yōu)化改進(jìn)
校園網(wǎng)云盤系統(tǒng)存在的安全問(wèn)題及對(duì)策
基于指紋身份認(rèn)證的固定通信臺(tái)站干部跟班管理系統(tǒng)設(shè)計(jì)
基于PKI與基于IBC的認(rèn)證技術(shù)比較
志愿服務(wù)與“互聯(lián)網(wǎng)+”結(jié)合模式探究
云計(jì)算與虛擬化
基于云計(jì)算的移動(dòng)學(xué)習(xí)平臺(tái)的設(shè)計(jì)
實(shí)驗(yàn)云:理論教學(xué)與實(shí)驗(yàn)教學(xué)深度融合的助推器
云計(jì)算中的存儲(chǔ)虛擬化技術(shù)應(yīng)用
Kerberos身份認(rèn)證協(xié)議的改進(jìn)
贺州市| 荆州市| 碌曲县| 乐亭县| 伊金霍洛旗| 黑龙江省| 北票市| 班戈县| 庐江县| 改则县| 娄底市| 襄汾县| 桃江县| 马边| 高邮市| 松溪县| 平舆县| 泸水县| 新巴尔虎右旗| 增城市| 苍溪县| 德格县| 蒲城县| 额敏县| 板桥市| 黄龙县| 东源县| 依安县| 永登县| 兴文县| 余姚市| 德令哈市| 华池县| 光泽县| 边坝县| 深州市| 随州市| 灵川县| 来凤县| 资溪县| 滨海县|