Recently, Virtual Reality has been increasingly researched on, especially there are a lot of studies on generating Virtual Fish on computer systems. In this paper, we propose an algorithm to automatically generate 3-D fish models from 2-D textures of fishes in illustrated books, handwriting, pictures, and so on. At first, We automatically separate several fins of fish from body using an original algorithm "Active Balloon". Next, users choose front view model and top view model of fish among the six samples each other. Finally, the number of patch is decreased without any influence on the accuracy of the generated 3-D model in an attempt to reduce the time cost when texture mapping is applied.
View full abstract