Multi algorithms for improving leukemia images edge detection technique

Edge detection is an essential pre-processing operation in image processing and pattern recognition. It involves identifying and tracing the sharp sudden discontinuities to extract meaningful information from an image. Edge detection simplify the analysis of an image by drastically reducing the amount of data to be processed and filtering out inadequate information, while at the same time preserving useful structural information about object boundaries in an image. The discontinuities signify the sudden changes in pixel intensity which describes boundaries of objects in a scene. The purpose of the present study is to detect the leukaemia edges in the white blood cell image. Toward this end, two distinctive procedures are implemented which are Ant Colony Optimization Algorithm and the gradient edge detectors (Sobel, Prewitt and Robert). The latter involves image filtering, binarization, kernel convolution filtering and image transformation. Meanwhile, ACO involves filtering, enhancement, detection and localization of the edges. Finally, the performance of the edge detection methods ACO, Sobel, Prewitt and Robert is compared in order to determine the best edge detection method which yielded optimal true edges of leukaemia in the white blood cell image. The results revealed distinctive results whereby the Prewitt edge detection method produced optimal performance for detecting edges of leukaemia cells with a value of 107%. Meanwhile, the ACO, Sobel and Robert yielded performance results of 76%, 102% and 93% respectively. Overall findings indicated that the gradient edge detection methods are superior to the Ant Colony Optimization method.