AI seminar: Image processing and the 2D Stockwell transform

Speaker: Allan Caine

In this talk, we will present the 2D Stockwell transform (2D s-transform) and its application to image processing. The 2D s-transform will be compared to the 2D Fourier transform. We will show that the 2D s-transform gives resolution in space while the 2D Fourier transform does not. To understand why this difference exists between the two transforms, we will look in detail at the Stockwell and Fourier transforms in their 1D versions. We will show that the Stockwell transform is a function of space and frequency; the Fourier transform is a function of frequency only.

In its 2D version, the Stockwell transform is a function of frequency in x and in y. These two parameters can be adjusted independently of each other. By varying the values of the two parameters, we can draw meaningful conclusions about the image under investigation.

We will conclude this talk with a summary of future directions in this exciting area of research.