Explore chapters and articles related to this topic
Computer vision
Published in Janet Finlay, Alan Dix, An Introduction to Artificial Intelligence, 2020
The Gaussian filter is a special smoothing filter based on the bell-shaped Gaussian curve, well known in statistics as the “normal” distribution. One imagines a window of infinite size, where the weight, w(x, y), assigned to the pixel at position x, y from the centre is w(x,y)=1πσ2exp[−(x2+y2)/2σ2]
Surface Features
Published in Wolfgang Osten, Optical Inspection of Microsystems, 2019
The most widely used filtering for removing noise from surface topography is the Gaussian filter [14], which can be defined as the convolution of the surface data with a Gaussian kernel. A Gaussian filter can effectively remove the speckle noise, but it will blur the edge. Mean and median filters [21] are widely used as noise-removing filters. The mean filtering is simply to replace each point value in a surface with the mean value of its neighbors, including itself. The mean filtering is usually thought of as a convolution filter; thus, it will also blur the edge feature. The median filter considers each point in the surface in turn and looks at its nearby neighbors to decide whether or not it is representative of its surroundings. Instead of simply replacing the point value with the mean of neighboring point values, it replaces the point value with the median of those values. It has been proved that for small to moderate levels of Gaussian noise, speckle noise, and salt-and-pepper noise (measurement outliers), the median filter is demonstrably better than the Gaussian filter at removing noise while preserving edges.
Deep Learning for Retinal Analysis
Published in Ervin Sejdić, Tiago H. Falk, Signal Processing and Machine Learning for Biomedical Big Data, 2018
Henry A. Leopold, John S. Zelek, Vasudevan Lakshminarayanan
This section highlights key image processing techniques often used during preprocessing and image analysis, some examples of which may be found in Figure 17.14. Gaussian filter functions, like those discussed in Section 17.2.2, are a prominent underpinning of many image normalization and edge detection techniques, such as image blurring or Canny edge detectors. There are a myriad of preprocessing techniques that can be used in many different sequences, including colour normalization, deblurring/contrast enhancement, edge detection, morphological reconstruction, and phase transformations. The most common methodology begins with single channel normalization through the application of histogram and/or median filters; which channel depends on the morphology of interest. The following steps usually focus on enhancing the image contrast with a combination of deblurring and adaptive edge enhancement techniques. For more information on retinal image normalization, please refer to Ref. [59].
A framework for fast automatic image cropping based on deep saliency map detection and gaussian filter
Published in International Journal of Computers and Applications, 2019
Ziaur Rahman, Yi-Fei Pu, Muhammad Aamir, Farhan Ullah
In image processing, Gaussian filter is used to blur or smooth the image. The Gaussian filter is a non-uniform low-pass filter, and it is based on the equation of Gaussian, can be used to generate a kernel. Normally, the kernel is an array that is used in convolution, convolution is involved in the multiplication of a set of pixels from the image with corresponding pixel values of the array in the form of convolution mask, has size, and the coefficient reduces with growing distance from the hub of the kernel. Different size kernel mask array contains a different pattern of numbers because one of the property of Gaussian distribution is that, it will be non-zero everywhere and thus, would need an infinitely big kernel. We use this method to blur the edges of the salient object. For this purpose, we get Gaussian filter using the following function.
Building a smart dynamic kernel with compact support based on deep neural network for efficient X-ray image denoising
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2022
Zouhair Mbarki, Amine Ben Slama, Hassene Seddik, Hedi Trabelsi
-The width, and the degree of smoothing of a Gaussian filter is parameterised by its standard deviation , and the relationship between and the degree of smoothing is very simple. A larger implies a wider Gaussian filter and greater smoothing. Engineers can adjust the degree of smoothing to achieve a compromise between excessive blur of the desired image features (too much smoothing) and excessive undesired variation in the smoothed image due to noise and fine texture (too little smoothing).
Brain tumor classification based on deep CNN and modified butterfly optimization algorithm
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Vinodkumar Jacob, G V R Sagar, Kavita Goura, P S Subhashini Pedalanka
The pre-processing is done to eliminate the irrelevant noises that exist in the images. Here, the Gaussian filter is adapted to perform the pre-processing, and its input processing is termed as , which is utilised for further processing. The advantage of using a Gaussian filter is that it removes the image’s noise and smoothest the input image to attain the practical BTC. Moreover, the Gaussian filter can eliminate the salt and pepper noise present in the images, and the Gaussian filter is expressed as,