Computer-Aided Diagnosis of Acute Lymphoblastic Leukaemia

Table 1

Advantages and disadvantages of different preprocessing methods.

Method	Advantages	Disadvantages
Histogram equalization	Simple technique to enhance the contrast of an image by utilizing its histogram. Useful for images having darker or brighter background and foreground.	It is only useful if the input image has low contrast; otherwise it can reduce the image quality. It cannot discriminate between noise and actual image, which can increases the contrast of noise.
Linear contrast stretching	Enhance the contrast of image by extending dynamic range of intensity values. Useful for low contrast images.	Main disadvantage of linear contrast stretching is that it is vulnerable to noise. An image having single outlier pixel can reduce the effectiveness of this operation.
Median filter	Popular nonlinear filter for removing salt and pepper noise. Useful for preserving sharp edges of image.	Only useful for images having low density noise. It cannot perform well for the images having high percentage of salt and pepper noise.
Minimum filter	Easy to implement. Useful for removing salt noise from the images.	Can remove image detail if images are highly noisy.
Gaussian filter	Useful for removing blur and noise from the images.	If used alone can blur the edges and reduce the contrast of images.
Unsharp masking	Simple method for image sharpening. Remove blurriness from the image.	Highly sensitive to noise because of linear high pass filter.