Most analysis whether they are analyzing data from a clinical trial or whether they are trying to analyze images. Similarly, analysis of data requires separation of noise from the signal. In imaging one of the newer techniques of segmenting images is called Multicut. This is best explained in an article that was written for the ilastik framework to analyze images.
While there also check out the video to do real time segmentation of tracking objects or animals.
{youtube} gQUrABQo6ow{/youtube}
These same methods may be applicable to analyzing complex data sets is two dimensions since the method has the ability to sort through neighbors. It would be great to see other applications of multicut to analyze complex data sets that are NOT images.