Quantcast
Channel: Machine Learning
Viewing all articles
Browse latest Browse all 62908

Why do texture filter banks change the size of the data?

$
0
0

I'm new to texture analysis. I was experimenting with some pretty standard filter banks when I came across something strange.

The data I'm working with is a 2D matrix of dimensions: 379x422.

When I put the data through the filterbank, the dimensions of the responses were 427x470.

The filterbank I'm working with is the The Leung-Malik (LM) Filter Bank. It consists of first and second derivatives of Gaussians at 6 orientations and 3 scales making a total of 36; 8 Laplacian of Gaussian (LOG) filters; and 4 Gaussians.

Essentially, I'm just wondering why the responses have more data than the original? What's going on here?

submitted by logrech
[link][4 comments]

Viewing all articles
Browse latest Browse all 62908

Trending Articles