Edge Detection Of Leukemia Cell Image Based On Spiking Neural Network

Because of the shortcomings of the traditional leukemia diagnosis, in this thesis, the separation of leukemia cells from background, and the segmentation of leukemia cell images are studied with digital image processing technology and other relevant knowledge by using Matlab, for the final purpose that the segmentation results can be used as a crucial step in the computer automatic diagnosis of leukemia.First, image preprocessing, there are two methods to use to separate leukemia cells from background, spatial domain processing method and frequency domain processing method. Secondly, introduced a lot of traditional image segmentation algorithms, such as thresholding, region-based segmentation, edge detection, segmentation using the watershed transform, etc.. Then use these methods to segment leukemia cell images, and analysis their advantages and disadvantages. Finally, the structure of spiking neural network (SNN) is presented, which is regarded as the third generation of ANNs and established based on the simulation of network model for biological visual cortex. Also the conductance based integrate-and-fire spiking neuron model is given, and then the neural network is applied to edge detection of leukemia cell images.Image preprocessing has been done to separate leukemia cells from background. Spiking neural network is used for the edge detection of leukemia cell images in internal for the first time. The experimental results show that edge detection based on SNN can be applied to leukemia cell images segmentation.