Object Tracking

Part 1 presents the methods based on block matching and object identifi cation using color and spatial information (Chapter 1). Chapter 2 presents methods that use the structural similarity measure in deterministic and probabilistic approaches based on hybrid particle filter. There adaptive tracking method is proposed in Chapter 3, which integrates online semi-supervised classifi cation and particle fi lter. Local features and global appearances and shape of dynamic objects methods are described in Chapter 4.Chapter 5 covers local and global tracking using the mean-shift and covariance matching.Fusion of confi dence maps in the visual object tracking procedure is described in Chapter 6. Solving the occlusion problems by coupling between optical fl ow and predictive algorithms using fuzzy control system is shown in Chapter 7.

Chapter 8 and Chapter 9 in Part 1 present the application of commercial soft ware and hardware solutions for object tracking. Chapter 8 presents a method for extraction and indexing of moving objects in H.264/AVC bitstream domains and Chapter 9 presents real-time implementation of graph cuts calculation on GPUs using CUDA for simultaneous object tracking and segmentation.

Part 2 presents descriptions of the application of automatic object tracking to some special purposes. The frameworks targeted for digital video surveillance tasks in urban environment and over wide-areas such as an airport, downtown of a large city or any large public area are presented in Chapters 10 and Chapters 11. Automatic calibration method for distributed sensors by using mobile robots is presented in Chapter 12. The method for tracking surfaces which deformations are diffi cult to be described, such as drapery of textiles, facial expressions or complex organ motion in medical image sequences is presented in Chapter 13. Chapter 14 describes method for coronary artery segment tracking in angiographic images.