Robotic Vision: Technologies for Machine Learning and Vision Applications

In this page we make publicly available a demonstration of a watershed Image Segmentation method based on a hybrid distance which is inspired in the selective behavior of the human vision system.

The method has 2 parameters (a,b) that specify the mixture model controlling the of activation of each component of the hybrid distance for the hybrid gradient. Besides, the method allows the specification of a decision threshold controlling the sensitivity of the method the region mergin.

For a fast test of the parameters (a,b, threshold). We shared this light application for Windows platforms [media: Hybrid_Watershed_Image Segmentation.zip]

The Sources. This class in C# and contains some methods for image segmentation based on the spheric approach.

For any question, fell free to ask me.

Platform support: This software has been developed in the .Net platform.