Hi,

              Thanks for reply . I have read the modules of scikit-image and understand the modules . First i was thinking of implementing otsu's thersholding algorithm .But it is already implemented . So now I am thinking of implementing multi-level thresholding or graph cut segmentation , I have read the graph based image segmentation paper . 

and also started implementing it on graph (currently ) , I have installation problems (with latest source code) due to some dependencies of  scipy on BLAS and LAPACK ,that i will resolve after my end semester exam(after 19th april) .

              and i also there is only canny edge detection algorithm for edge detection . although it is better the many other edge detection algorithm . So is it useful if there exist some more edge detection techniques ?  

             and out of these above three which is more important to implement ( according to me since segmentation is more difficult problem so segmentation should be important ).