Lavatube - A Software Framework for Computer Vision Research and Development
Kenji Iwata, Yutaka Satoh, Katsuhiko Sakaue · 2009
We develop Lavatube as a software framework for efficient research, development, and education. Lavatube is an object-oriented framework optimized for constructing a computer vision system, particularly a video and image processing system. Lavatube enables a description of a processing extension by combining various functional components. Since the data flow is easy to describe by graph-connecting icons on a GUI, a system can be created intuitively. The standard support of input from a USB camera and various video files makes complicated work or knowledge unnecessary. For efficient and clear-cut system construction, Lavatube also provides functions for the dynamic generation of parameter setting dialog boxes and for perpetuation by XML. Some actual video processing cases as examples are also introduced. 1.