Design and implementation of high speed Gabor filter with variable thresholding process for disease detection

Abstract

In current scenario, there are number of skin related diseases such as tonsillitis, tumors, skin cancers, etc… which can be detected at an early stage and can be cured. For this a new idea is proposed which aims at designing a Gabor filter that makes accurate detection of diseases. We use an efficient pipelined architecture for input. Image segmentation for the detection of diseases is done using Gabor filter that in turn uses shift-add CORDIC algorithm for making the designing of Gabor filter easier. In addition to this algorithm, we use a new algorithm called Variable Thresholding algorithm that sets a threshold value during the decision making period. The algorithm is named variable because, the threshold value can be varied depending upon the diseases we are going to detect. All these algorithms are written in VHDL Language in MODELSIM software, which makes the designing and decision making process more effective. This technique is also implemented on FPGA/CPLD kit that shows us the presence of disease manually through lighting of LED's. Thus the decision of disease is made more accurate and faster through this design.