|M.Sc Thesis||Department of Computer Science|
|Supervisor:||Prof. Gotsman Chaim Craig|
There are many possible two-dimensional views (or images) of a given three-dimensional object and most people would agree that some views are more aesthetic and/or more “informative” than others. It would be very useful, in many applications, to be able to automatically compute these “good” views.
Although all measures of the quality of a view will ultimately be subjective, hence difficult to quantify, we believe that there are some general computational principles that may be used to address this challenge.
Previous research on this topic dealt with only some aspects of this question. In our work we review these results and then present a general analysis of the problem. We propose a two-phase method for the computation of a “good” view. The first phase—view filtering—performs a global view-independent analysis of the object in order to generate a small number of candidate views. In the second phase we apply a view-dependent measure—a view descriptor—to select the most “informative” view among these candidates. For both phases, a number of heuristic methods were examined. We elaborate on these and show results.