Sharma, Ravi Karan and Nibariya, Rohit (2009) Study Of Gaussian & Impulsive Noise Suppression Schemes In Images. BTech thesis.
Noise is introduced into images usually while transferring and acquiring them.The main type of noise added while image acquisition is called Gaussian noise while Impulsive noise is generally introduced while transmitting image data over an unsecure communication channel , while it can also be added by acquiring. Gaussian noise is a set of values taken from a zero mean Gaussian distribution which are added to each pixel value. Impulsive noise involves changing a part of the pixel values with random ones. Various techniques are employed for the removal of these types of noise based on the properties of their respective noise models. Impulse Noise removal algorithms popularly use ordered statistics based ¯lters. The ¯rst one is an adaptive ¯lter using center-weighted median. In this method, the di®erence of the center weighted mean of a neighborhood with the central pixel under consideration is compared with a set of thresholds. Another method which takes into account the presence of the noise free pixels has been implemented.It convolutes the median of each neighborhood with a set of convolution kernels which are oriented according to all possible con¯gurations of edges that contain the central pixel,if it lies on an edge. A third method which deals with the detection of noisy pixels on the binary slices of an image is implemented. It is based on threshold Boolean ¯ltering. The ¯lter inverts the value of the central pixel if the number of pixels with values opposite to it is more than the threshold. The fourth method has an e±cient double derivative detector, which gives a de- cision based on the value of the double derivative. The substitution is done with the average gray scale value of the neighborhood. Gaussian Noise removal algorithms ideally should smooth the distinct parts of the image without blurring the edges.A universal noise removing scheme is implemented which weighs each pixel with respect to its neighborhood and deals with Gaussian and impulse noise pixels di®erently based on parameter values for spatial, radiometric and impulsive weight of the central pixel. The aforementioned techniques are implemented and their results are compared subjectively as well as objectively.
|Item Type:||Thesis (BTech)|
|Uncontrolled Keywords:||Noise Suppression, Threshold Boolean Filtering, Switching Median Filter, Universal Impulse Detector, Center-Weighted Median Filters,|
|Subjects:||Engineering and Technology > Computer and Information Science > Image Processing|
|Divisions:||Engineering and Technology > Department of Computer Science|
|Deposited By:||Rohit Nibariya|
|Deposited On:||22 May 2009 16:23|
|Last Modified:||14 Jun 2012 10:29|
|Supervisor(s):||Sa, P K|
Repository Staff Only: item control page