It is appreciated based on the foregoing discussion that in some embodiments, the method may include different or additional steps, such as retrieving the centroid vector of each cluster and/or the distances between each two clusters, and using the centroid vector of each cluster and/or the distances between each clusters to construct a graph, and performing a graph-based image processing task based on the graph. As discussed above, in some embodiments, the graph-based image processing task may include a segmentation operation that can use the graph and one or more user annotations to segment the biological specimen image into a plurality of segments. The method may also include determining a set of different weights for the different image features, where the generation of the predefined number of clusters and the association of each superpixel with the cluster are based at least in part on the set of different weights.

In the foregoing discussion, various devices, engines, units, or blocks (e.g., some or all blocks system 100) were described as being implementable using hardware, software, firmware, or any combination thereof. It must be appreciated by a person skilled in the art in view of the foregoing discussion that the hardware may include any type of analog and/or digital circuitry, such as integrated circuits (IC) (e.g., application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs)), or any other type of special- or general-purpose electronic circuitry. It must be further appreciated by a person skilled in the art that the software or firmware may include any type of processor executable instructions that can be stored on any type of tangible non-transitory computer-readable medium, where the instructions can be executed by a processing resource, causing the processing resource to implement the functionality of the respective component.