Analysis and evaluation of hard and fuzzy clustering segmentation techniques in burned patient images

摘要

For the pre-diagnosis of a hospitalised burned patient, it is necessary to estimate the extent of the injuries. At present, this process is carried out visually by the physician, and for that reason the management of the patient relies heavily on a clinician's experience. The design of an automatic system that allows for the classification and establishment of a treatment plan not only will add to the improvement of health assistance in the hospital, but can also be used by rural centres to decide which patients will need specialised care.With these objectives in mind, several algorithms for the segmentation of burned patient images were implemented and tested to assess their validity. Three algorithms were implemented as follows: the first one is a hard algorithm that was designed using a supercomputer to calculate its parameters; and the other two are based on different approaches using fuzzy algorithms. In this paper, these hard and fuzzy algorithms are described, together with the results obtained, so as to compare their performance.