Regions of interest selection in histopathological images using subspace and multi-objective stream clustering

The advances of deep learning in histopathology show the ability to assist pathologists in reducing workload and avoiding subjective decisions. Such algorithms lead to a more reliable diagnosis because they give computer-based second opinions to the clinician. However, in histopathology cancer image analysis, pathologists mostly diagnose the pathology as positive if a small part of it is considered cancer tissue. These small parts are called regions of interest or patches. Finding the relevant patches is crucial as it can save computation time and memory. Also, deep learning systems can receive only small inputs, and these patches represent the best input. This paper proposes a new clustering algorithm for the patch selection based on subspace clustering. This technique discovers clusters embedded in multiple, overlapping subspaces of high-dimensional data. The proposed algorithm manages to find the data’s best partitioning and the images’ relevant patches.