COMP61342 : Computer Vision

MSc Module part of Taught Postgraduate Programmes


Image smoothing and Sharpening

At the end of this practical you should be able to:
- understand the convolution in image processing.
- processing image using different filters.


Gaussian Smoothing
Difference of Gaussian

Matlab has provided a guide on performing convolution on images. Run "demos" command, and find the "User's Guide" for "Convolution" in "Image Processing Toolbox". (A topic under "Designing and Implementing 2-D Linear Filter for Image Data", and "Designing and Implementing Linear Filters in Spatial Domain").

Follow the guide, and construct a few different covolutions kernels for smoothing and sharpening images. Apply the filters on your own image.

Difference-of-Gaussian (DoG) is technique widely used in many Computer Vision applicaitons. Try if you can construct some results of DoG by applying Gaussian Filtering on your image at difference scales. (There are many references to DoG on the Internet).

Back to Self-Study

The last version of Matlab we checked with this tutorial available is R2012a.