Click&draw selection

In interactive shape modeling, surface selection is one of the basic, cornerstone interaction: this task may be performed hundreds of times by the user for the modification of a single shape. Despite the numerous automatic selection methods introduced in the literature, this remains a cumbersome operation in a number of scenarios, where the user is ultimately asked to paint over everything she wants to select. Without imposing a full visual grammar, we observe that a more effective selection process can be designed around a simple classification of the user interaction: point click, open and close strokes. Our basic idea is to relate this simple classification to a specific set of selection algorithms, targeting the three main classes of surface selections: connected components, parts and patches. We also address the problem of repetitive similar selections by providing an automatic expansion process which captures regions which are detected as similar to the selected one.