透過FDHI實現人類步伐辨識; Human identification using Gait Features via Forward Difference History Image

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/48502

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/48502

Title:	透過FDHI實現人類步伐辨識;Human identification using Gait Features via Forward Difference History Image
Authors:	賴宣憲;Syuan-Sian Lai
Contributors:	資訊工程研究所
Keywords:	步伐辨識;FDHI
Date:	2011-08-16
Issue Date:	2012-01-05 14:56:33 (UTC+8)
Abstract:	生物認證一直都是相當熱門的議題，其研究成果可運用在安檢系統、ATM的認證或許多的商業應用系統認證上。舉凡指紋、人臉、虹膜都可用來當作生物特徵。但以上的各種特徵都必須在固定的環境下辨識，而用人類的步伐來當作生物特徵是近幾年新興的議題，好處就是不需在固定的環境下也能取得正確的特徵以及良好的辨識率。現有的人類步伐特徵擷取方法舉凡GEI、GHI、GMI…等，都是將一整個步伐週期結合為一張影像當作特徵，而本論文探討人的步伐在一個週期下是否能再細分成四個週期並且分別結合為一張影像當作特徵，並透過不同的結合方法，發展出一套自己的步伐認證方法，也就是Forward Difference History Image(FDHI)。在實驗的部份，我們將比較GEI以及FDHI的辨識率，並透過實驗結果說明在四個週期下的特徵個別效果；我們比較了GEI以及FDHI，並做了大量的實驗，根據辨識率上的比較，可以得知FDHI優於GEI。 Human identification is an important issue in identity authentication which can be applied in many applications, such as security monitoring system, ATM authentication, and personal authentication in businesses transactions. There are many mature image-based human identification techniques that have been developed, such as fingerprints, face, and iris biometric modalities. However, these methods impose severe constraints, such as requiring of a cooperative subject, views from certain aspects, and physical contact or close proximity. To relieve these constraints, human gait identification is a new choice to remedy the problems. The existing human gait identification methods, including GEI、GHI、GMI…etc, are formed by combining the whole human gait cycle into one image. In this thesis, an effective human gait identification method is presented by separating one cycle into 4 cycles via different combination methods. Experimental results reveal the feasibility and effectiveness of the proposed method in gait identification. We also compare the performance of GEI and FDHI and through experimenting to explain the effects of the four different cycles. The results confirm that our proposed FDHI identification is better than GEI identification.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	607	View/Open

社群 sharing

Loading...