Do you guys know image processing? Humus?

K.I.L.E.R

Retarded moron
Veteran
Is this a lot?

Image acquisition and image files. Perception; human and computer vision. Sampling and quantization. Point processing: thresholding and histogram methods. Spatial filtering: linear and non-linear filters; convolution and correlation, adaptive filtering. Application to noise reduction and edge detection. The Fourier transform and its applications to periodic noise removal, de-blurring and inverse filtering. Hough and distance transforms. Mathematical morphology: shape analysis, shrinking, thickening and skeletonization. Image coding: run-length encoding and Huffman coding; applications to compression. Colour processing: colour spaces and colour space conversion.

http://csm.vu.edu.au/courses/subjects/rcm2511
 
I went through the equivalent of this, and it was also done with Matlab :)
it's not that much stuff, and not hard, even quite fun as you'll do some photoshop-like stuff with simple code. (but later requiring a bit of thought).

Matlab is a language with simple and very powerful handling of arrays/matrices (then it's plain classic structured programming like C); its functions load images and do the Fourier transform for you, and easily plot the data. but you do the other stuff yourself.

you should be familiar with quite a number of things. "Sampling and quantization" is about resolution and bpp, colour spaces is about RGB, YUV, hue/saturation/luminosity etc.
Thresholding and histogram methods, this is easy.
Most annoying is applying filters, dithering and morpho maths by hand for the examination ;)
 
Will this give me the knowledge of doing HDR?
I know the steps to do HDR.
I can figure out how to do guassian interpolation and grayscaling on my own as long as I know how the stuff behind it works.

I know about sampling and avraging values of colours and stuff.
I used that in OpenGL to apply pre-processed ray tracing effects using buffers and lots of masks.
 
Sounds good. I had a very similar course while I was at university. None of this is really complex (except maybe for the Fourier transform, but you'll only have to understand it, not do it yourself). I think you'll enjoy it.

K.I.L.E.R said:
Will this give me the knowledge of doing HDR?
I know the steps to do HDR.
So why do you want to learn what you already know?
 
Xmas said:
None of this is really complex (except maybe for the Fourier transform, but you'll only have to understand it, not do it yourself).

This is not complex from a conceptual point of view, but to implement this in an efficient way can be quite challenging especially if you intend to do it with programmable logics (such as an FPGA). Even a simple convolution can be quite something if you want optimal performance.
 
Image acquisition and image files. Perception; human and computer vision. Sampling and quantization. Point processing: thresholding and histogram methods. Spatial filtering: linear and non-linear filters; convolution and correlation, adaptive filtering. Application to noise reduction and edge detection. The Fourier transform and its applications to periodic noise removal, de-blurring and inverse filtering. Hough and distance transforms. Mathematical morphology: shape analysis, shrinking, thickening and skeletonization. Image coding: run-length encoding and Huffman coding; applications to compression. Colour processing: colour spaces and colour space conversion.

Add page numbers, and you've got the table of contents for the book by Gonzalez and Woods, which they appear to be using. Compared to books from other courses I've had, it's one of the most interesting. Have had no use for the information ever since, though...

Ditto to what others said. Also, the Matlab assignments on that course were some of the most interesting ones. Sure beats control systems simulations in coolness factor for a newbie.
 
Well the other day you were asking what the use of fourier transforms was. This will show you some of them.

Image processing was one of the more interesting courses I did at uni.

CC
 
K.I.L.E.R said:
What can this help me with besides messing around with shaders?

Well, messing around with shaders is a good use of it. Other than that I guess there are industrial applications of it (textbook example: analyze camera images of objects passing by on a conveyor belt), but if you ever plan on working with graphics a course like this is definitely great to have. I'll have to agree it was one of the most interesting courses I read at uni.
 
Is it possible to decode a heatvision video frame by frame of a man making out with a donkey like the one I posted in the P&R forum into colour?
 
Humus said:
if you ever plan on working with graphics a course like this is definitely great to have. I'll have to agree it was one of the most interesting courses I read at uni.
Same for me. I am still working on one of the assignments 5 years later (my image compression)... ;)
 
Back
Top