December 2014
Beginner to intermediate
154 pages
4h 7m
English
As introduced in the previous chapter, a large group of views not only provide rich information but also produce redundancy. To overcome this limitation for a given group of initial captured views, it is important to select representative views from a pool of views. The objective of view selection is to determine a small set of compact but discriminative views for 3-D object representation. Existing view selection methods can be generally divided into two categories: unsupervised methods and interactive methods. In unsupervised methods, a common procedure is to conduct view clustering to select representative views from each view cluster. Interactive methods, on the other hand, take into account user relevance ...