Face detection and gesture recognition for human-computer interaction (Q2772095)

This book is concerned with vision-based interfaces between man and machine. NEWLINENEWLINENEWLINEThe first part of this book aims to extract two-dimensional motion across image frames and classify underlying three-dimensional motion patterns. The authors developed a method for extracting and classifying two-dimensional motion in an image sequence based on motion trajectories. For concreteness, they focus on image sequences showing hand motions for signs in American Sign Language. The same method can be adapted to recognize other motion patterns such as walking and running.NEWLINENEWLINENEWLINEIn this context, it is important to identify certain body parts of the signer in an image sequence. Also, it is important to locate the hand position relative to head and the rest of the body. Therefore, a related problem is to detect human faces robustly. In the second part of this book, four face detection methods are presented to detect human faces in color and gray-scale images. Among these, two of them use multimodal density functions and one uses the SNoW learning algorithm to detect faces in gray-scale images. For color images, the authors develop one face detection method based on the characteristics of skin color and facial structure. The motivations behind these methods and their difference are also discussed.NEWLINENEWLINENEWLINEIn the last part of this book, a SNoW-based face detection method in still images is presented. To quantify and explain the experimental results, a presented theoretical analysis shows the advantage of this architecture and traces it to the nature of the update rule used in SNoW.NEWLINENEWLINENEWLINEAn comprehensive review on face detection is presented in Chapter 2. After analyzing these algorithms and identifying their limitations the authors conclude with several promising directions for future research.NEWLINENEWLINENEWLINEThey then present their algorithm for extracting and recognizing motion patterns in Chapter 3. Experimental results on extraction and recognition of motion patterns associated with hand gestures using a set of 40 ASL signs are presented. In Chapter 4, a skin color model using a mixture of Gaussians is described. Two algorithms using multimodal density functions are introduced in Chapter 5 to detect faces in gray-scale images. Experimental results on several benchmark databases are presented. In Chapter 6, a SNoW-based face detector is described and explained why and when SNoW performs well in the face detection problem. Chapter 7 is the conclusions and comments on future research.