Cris’s Image Analysis Blog

Hollywood distorts everything, but especially likes to make up its own science. People writing the scripts for these movies clearly have no understanding of even basic science. They read too much stuff on the Internet about quantum mechanics, radiation, genetics, you name it. They then proceed to make up their own explanations for what they read, and give these explanations to their audience. The biggest propaganda machine the world has ever known is working hard to misinform the world about science.

Read the rest of this entry »

If you don’t use DIPimage, you probably use MATLAB’s Image Processing Toolbox. This toolbox makes it really easy to do convolutions with a Gaussian in the wrong way. On three accounts. The function fspecial is used to create a convolution kernel for a Gaussian filter. This kernel is 2D. That’s the first problem. The other two problems are given by the default values of its parameters. The default value for the kernel size is [3 3]. The default value for the σ (sigma) is 0.5. Read the rest of this entry »

I read a very nice article in the latest IEEE Signal Processing Magazine: Prandoni, P. and Vetterli, M., “From Lagrange to Shannon… and back: another look at sampling,” IEEE Signal Processing Magazine 26(5):138-144, September 2009. The authors make a case for teaching signal processing starting with discrete time, and then moving to continuous time. I don’t agree, but they expose their case very nicely. But I did learn something new from this paper, which is why I am writing this. It turns out that the Lagrange interpolation polynomials converge to the sinc function as the polynomial order goes to infinity.

Read the rest of this entry »