I am studying on image noise reduction. I have a doubt on image smoothing and denoising. Please anyone clarify the difference between image smoothing and denoising with suitable example.
Smoothing is one of the denoising methods. Smoothing basically acts as low pass filtering and high frequency components such as edges and boundaries are removed; rather replaced by either average or mean values of gray scale values of all the pixels. Suppose, due to poor lighting the image quality is not proper which could be enhanced by using any smoothing domain filters; whereas Denoising includes all image enhancement methods: in spatial as well as transform domain. I believe that if you google my name as Dr G R Sinha and image enhancement, you shall certainly get few research papers. You also can find these concepts in my two books: http://www.wileyindia.com/biometrics-concepts-and-applications.html
and http://www.amazon.in/Medical-Image-Processing-Concepts-Applications/dp/8120349024
Smoothing is one of the denoising methods. Smoothing basically acts as low pass filtering and high frequency components such as edges and boundaries are removed; rather replaced by either average or mean values of gray scale values of all the pixels. Suppose, due to poor lighting the image quality is not proper which could be enhanced by using any smoothing domain filters; whereas Denoising includes all image enhancement methods: in spatial as well as transform domain. I believe that if you google my name as Dr G R Sinha and image enhancement, you shall certainly get few research papers. You also can find these concepts in my two books: http://www.wileyindia.com/biometrics-concepts-and-applications.html
and http://www.amazon.in/Medical-Image-Processing-Concepts-Applications/dp/8120349024
Smoothing is basically a blurring process. It is frequently used in image processing. Denoising is about removing some unnecessary disturbances impregnated into an image for various reasons. Denoising is a challenging task when edges are to be strictly preserved.
Denoising is removal of unwanted signals ( random variations of color and brightness) using various spatial and transform domain filters.
Fundamentally , convolution is an important operator . You can take a mask and convolve for noise reduction or blurring. In denoising for eg: salt and pepper (common in images) median filtering is best suited because here you are not just convolving but non-linearly replacing . They(filter responses) are unfortunately noise based.
Smoothing is when there are high freq components (say due to aliasing or sampling ), you low pass filter (again convolve or anti-aliasing) resulting in smooth transitions -subjective quality of the image . Too much , say spread factor (sigma of Gauss) choice will blur the image . So alternates such as box filter/ Bartlett / piecewise /windowed exists.
Bottom- line: Operations may be same but filter design and response are critical.
Smoothing is an Enhancement technique for improving the image quality (contrast enhancement). It can use for clear vision or as preprocessing in various applications such as Medical Imaging , Satellite Imaging etc.. . While Denoising is the process of removing the noises from image (some time it is called Restoration ) For example noise because of capturing devices or machine or because of object movement or because of compressed storage or some time because of channel noise. Examples of noises are thermal noise ,blur,motion etc.. For developing a new denoising algorithm ,which required to test with some added noises while developing a smoothing filter , added noise testing it not necessary .
- Badges
- Science topic
- Similar topics
- Computer Science
- Images