Felzenszwalb, P. F. & Huttenlocher, D. P. Efficient graph-based image segmentation International Journal of Computer Vision, 2004, 59, 167-181
this is a method for over segmentation but object segmentation result could not be produce.
you can also refer to this article
Carreira, J. & Sminchisescu, C. Constrained parametric min-cuts for automatic object segmentation Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on, 2010, 3241-3248
this method is similar to Graphcut, but EM is not needed in this method. The comparision between neighbour region is achieved through a simple way instead of EM.
Segmentation methods depends on what what you want to segment from image.
load your image to see what algorithm fit with your request. some algorithms need to transform image to fit with it or if you do not how to transfer your image you need to know how the algorithm work so you can change in its rule to give you a perfect result after comparing the segment result with the ground truth.
Segmentation using geodesic distances proved to be very efficient because it does not involve any energy minimization step. When coupled with superpixels it becomes even more efficient [e.g. hundreds of segmentations per second in the "geodesic object proposal method"].
You can implement it on your own with few lines of code, following the description of the Dijkstra's algorithm. You can contact me for further details.
There are more than 40 segmentation methods. However, all depends on researcher choice, I usually use the Otsu method if the image shows a bi-modal structure but one of the most recommended is Mathematical Morphology developed at INRIA by Levy-Vehel et al.