Thermal Image Processing Approach to Detect Malaria using Fuzzy Logic

The thermal image processing technique for detecting malaria using General Fuzzy Min-Max neural network (GFMM). For detecting malaria, image should go through 4 standard steps, pre-processing, segmentation, feature extraction and selection and classification. Median filter is used in pre-processing step which reduces salt-and-pepper noise of the image. The filtered image is then segmented with the help of Otsu thresholding technique which automatically computes the optimum threshold partitioning the two classes such that spreading is minimal. The features of the segmented part are extracted by Gray Level Co-occurrence Matrix (GLCM), which extracts the infected part of the malaria blood cell. This matrix holds data of gray values of every pixel at its corresponding location. Finally, the GFMM is performed on the extracted data for classification. It performs classification along with clustering, which provides efficient way in recognizing and searching the infected part of the cell. General Terms Image pre-processing, Segmentation, Feature extraction.