Question

I am completely new to image processing. I know nothing about what is JPEG internally and how it works.

I wonder, if I can find somewhere piece of ruby code performing following simple operation:

  1. Open jpeg file.
  2. Iterate through each pixel and set it's color to fx green.
  3. Write result to another file.

I am especially interested in how this can be accomplished using ruby-vips library
https://github.com/ender672/ruby-vips

My goal - to learn how to perform basic image processing operations using ruby-vips (gamma correction, brightness, tint, ...)

Any links to working examples more complex than 'hello world'-like one on ruby-vips's github page would be highly appreciated!

If there are alternatives to ruby-vips, I would be thankful for them too.


UPDATE

Much has happened since I asked this question:

Was it helpful?

Solution

update ruby-vips has changed a bit since this answer was written. I've revised it for the current (2018) version.

I'm one of the maintainers of libvips, the image processing library that ruby-vips wraps.

Tim's ruby-vips repository hasn't been touched for a while. I have a fork here that works with current libvips:

https://github.com/jcupitt/ruby-vips

There are some examples here:

https://github.com/jcupitt/ruby-vips/tree/master/example

To set the red and blue channels to zero and just leave a green image you might multiply R and B by zero and G by 1. ruby-vips uses arrays to represent pixel constants, so you can just write:

out = in * [0, 1, 0]

A complete runnable example might be:

#!/usr/bin/ruby

require 'vips'

im = Vips::Image.new_from_file '/home/john/pics/theo.jpg'
im *= [0, 1, 0]
im.write_to_file 'x.jpg'

There's a trick you can use for new_from_file: if you know you will just be doing simple top-to-bottom operations on the image, like arithmetic or filtering or resize, you can tell ruby-vips that you only need sequential access to pixels:

im = Vips::Image.new_from_file '/home/john/pics/theo.jpg', access: :sequential

Now ruby-vips will stream your image. It'll run the load, the multiply and the save all in parallel and never keep more than a few scanlines of pixels in memory at any one time. This can give a really nice improvement to speed and memory use.

To change image gamma you might try something like:

im = im ** 0.5 * 255 / 255 ** 0.5

Though that'll be a bit slow (it'll call pow() three times for each pixel), it'd be much faster to make a lookup table, run the pow() on that, then map the image through the table:

lut = Vips::Image.identity
lut = lut ** 0.5 * 255 /255 ** 0.5
im = im.maplut lut

Any questions, please feel free to open them on the rubyvips issue tracker:

https://github.com/jcupitt/ruby-vips/issues

OTHER TIPS

I'm sorry I don't know ruby-vips, but ImageMagick is a classic when it comes to image processing. There are Ruby bindings in the form of RMagick (current repo), and you can derive a lot of functionality from the ImageMagick docs, but there are also three tutorials here, as well as a lot of examples on the web.

If you really want to go deep into the theory of image processing, which in its roots is a form of signal processing (this is totally exciting and rewarding as it often allows you to apply very similar algorithms on images and audio/video signals, but it will ultimately get very heavy on math - Fourier transforms), then, if mathematics don't scare you, I can only recommend to read the book by Gonzalez and Woods, I would say it's the definite reference in this field. It's expensive, but there's all you need in there to get you started and well beyond. Here's also a page with links to free ebooks if you would like to get started without spending lots of money first.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top