Digital Images Research Paper

3541 Words15 Pages

INTRODUCTION Chapter 1

A digital image is a numeric representation of a 2-dimensional image. It is represented as a finite set of some digital values, which are picture elements/pixels. Pixel values represent grey levels, colors, intensities, heights etc. Also, a digital image is an estimation of a real scenario, which is explained by digitization.
Digital Image Processing, as name expresses, is processing applied on digital images. It includes a number of techniques which are used to manipulate the digital images by computers. The question arises is that what is the need of processing the digital images? The answer is because the image received from the sensors on the satellite can contain some flaws or deficiencies like …show more content…

Therefore, presence of these characteristics makes retinal images complex to understand. Intensity of retinal images is also an issue to be considered because there is less difference in the foreground and background intensities in these images. If we considers the grey level images then we observes that the shape, size and local grey level of blood vessels may vary hugely and many background features can have similar attributes similar to vessels. Signal noise, drift in the intensity of an image and also, lack of contrast in image also pose tough challenges to the extract blood vessels in the retinal images.
In this thesis, a major hindrance in understanding the retinal images, noise, is considered. The process of denoising is a crucial and significant area that is strongly needed to be performed to understand the retinal images better. Noise in Digital Image …show more content…

e^(-az)@0 f or z<0)┤ forz≥0 μ=b/a σ^2=b/a^2
K=(a(〖b-1)〗^(b-1))/(b-1)! e^(-(b-1))
1.3.5. Rayleigh noise
Velocity images and radar range and velocity image generally contain noise of Rayleigh distribution. Figure 1.8: Rayleigh Noise (Kamboj and Rani)
Probability density function, Mean, Variance of Rayleigh noise are respectively: p(z)={█(2/b (z-a) e^(〖-(z-a)〗^2/b)@ 0 for z<a)┤ for z≥a μ=a+√(πb/4) σ^2=(b(4-μ))/4
After a brief discussion about noise, it is observed that noise is an unwanted signal but, this is not true for all the noise, that is, noise is always not bad (which can be a special case). But generally, it is important to remove noise because noise is visually unpleasant, it is also bad for compression and for analysis. Presence of noise in an image provides low accuracy to analyze the image. Noise removal can be done in a number of ways. In this thesis, noise is going to be removed through ICA(Independent Component Analysis) which will separate the noise signal and that will be discussed in the chapter 3- Materials and Methods.
In the next section, the very important key stages in digital image processing is discussed that is Image

More about Digital Images Research Paper

Open Document