Externally indexed torrent

If you are the original uploader, contact staff to have it moved to your account

Textbook in PDF and DJVU formats

A few major edited books treat face recognition. Among them are the first and seminal "Face Recognition: Prom Theory to Applications" (Wechsler et aI., Springer, 1998), and most recently the "Handbook of Face Recognition" (Li and Jain, Springer, 2005). This book is the first to comprehensively address the face recognition problem in its entirety, while drawing inspiration and gaining new insights from complementary fields of endeavor, such as neurosciences, statistics, signal and image processing, computer vision, machine learning and pattern recognition, and statistical learning. The various chapters treat topics related to how people represent, process and/or respond to the human face, modeling and prediction, the face space, identification and verification, face detection, tracking and recognition, 3D, data fusion, denial and deception using occlusion and disguise, performance evaluation and error analysis, and finally, competing security and privacy considerations.
The underlying theme of the book is that the biometric inputs chart continuous and coherent space and time manifolds, which facilitate their recognition. Face recognition is dynamic rather than static. It continuously iterates, making specific interpretations and assigning confidence to them. Supporting and non-accidental evidence is accrued in an active fashion, leading to lower uncertainty in the recognition decisions made, and resolving ambiguity, if any. Integral to face recognition are advances in pattern recognition. Novel methods are proposed here to handle real life applications where variability, incomplete, noisy, distorted and/or disguised patterns are usually the norm rather than the exception. The overall goal of the book is applied modem pattern recognition, with the understanding that the novel methods described here apply to any objects. The face pattern is only one of the object patterns that surround us and need to be recognized. The scope for pattern recognition (Rosenfeld and Wechsler, 2000) is much wider here because among other things both training and testing can take place using incomplete or camouflaged/disguised patterns drawn from single and/or multiple image sets