Optimal Superpixel Kernel-Based Kernel Low-Rank and Sparsity Representation for Brain Tumour Segmentation

Algorithm 2

The algorithm of the OSK-KLRR-SR classifier.

Input	: Multimodal brain tumour MR fusion data ; training sample set and corresponding label set.
Output	:Label set of the testing samples .
	Step 1: Initialize .
	Step 2: Compute the optimal superpixel kernel using Algorithm 1.
	Step 3: Compute the matrixes with elements and with elements .
	Step 4: While (oror) do
	Step 5: Update .
	Step 6: Update .
	Step 7: Update .
	Step 8: Update , ,
	Step 9: Update , where and .
	Step 10: Update .
	Step 11: End while
	Step 12: Compute the residuals of the sample belonging to each category , where .
	Step 13: Determine the category of the sample .