Interactive Texture Segmentation using Random Forests and Total Variation

Hypothesis The segmentation quality depends on a strong description for F and B. In order to model hypotheses based on different high-level features, we need an efficient learning algorithm capable of handling arbitrary input data. Random Forests (RFs) are fast to compute while yielding state-of-the-art performance in machine learning and vision problems. Their parallel structure dedicates them to GPU implementations. Recently, an online version of RFs has been proposed [2], which renders retraining of the whole forest upon additional user input unnecessary.