Files & Code: Active Contours Image Segmentation Based on Bhattacharya with Optical Flow

Files

Gray-scale(One-dimensional)
Not the main model, but good for learning purposes.
Color Features (Two-dimensional)
Not the main model, but very close. It uses color features.
Motion Features (Two-dimensional)
The main model, it works on two frames of a moving object.

The code is based on Dr. Oleg Michailovich paper, and his demonstration code, on Bhattacharyya Gradient Flow. [Michailovich, O. Rathi, Y. Tannenbaum, A. "Image Segmentation Using Active Contours Driven by the Bhattacharyya Gradient Flow" 2007]
In calculating the optical flow, we used Thomas Brox's Large Displacement Optical Flow method. [T. Brox, J. Malik. "Large displacement optical flow: descriptor matching in variational motion estimation" 2011]
In the case of segmentation based on color features, the optional projection of RGB to two-dimentional space is performed using Fast ICA algorithm by Aapo Hyvarinen. [A. Hyvärinen. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis. 1999]