傳統(tǒng)的身份驗(yàn)證主要有帳戶/密碼或智能卡等方式，存在著易被破解盜用和遺忘丟失
等風(fēng)險(xiǎn)。該文提出了一種在遠(yuǎn)程登錄過(guò)程中利用人臉識(shí)別技術(shù)進(jìn)行身份驗(yàn)證的解決方案。該方案在客戶端進(jìn)行人臉檢測(cè)和初步特征提取，由服務(wù)器進(jìn)行特征匹配和身份驗(yàn)證。其中人臉特征提取采用了奇異值分解和主分量分析法，身份驗(yàn)證則采用了以類(lèi)內(nèi)平均距離為判據(jù)的算法。實(shí)驗(yàn)證明了該方法的有效性。
關(guān)鍵字：人臉識(shí)別；特征提??；特征匹配；遠(yuǎn)程登陸；身份驗(yàn)證；
Abstract: The traditional authentication takes mainly use of account/password or smart card. But this method has some disadvantages such as that it can be stolen and decrypted. So this paper presents a solution which using face recognition technology for identity verification in remote login. In this subject, it detects the face and extracts the features at client and matches the features and verifies the identity at server. Here, we use the singular value decomposition and principal component analysis for facial feature extraction, using the average distance category as discrimination on the basis of authentication. The experiment shows that the method has a good performance.
Key words: face recognition; feature extraction; feature matching; remote landing; authentication