1. 19
  1.  

  2. 8

    I used libvips back in 2008-2010 to write software for analyzing medical microscopy images! In the lab I worked in, we literally couldn’t afford enough RAM to load entire multi-gigabyte images into memory all at once. Libvips let us work mostly from the images on disk with little RAM and implement registration, color correction and such. Now that everyone has 32GB in their pocket, memory usage seems like a completely silly consideration. But it was vital back then!

    Nip2 (a spreadsheet-like image editor on top of libvips) is a really neat idea, and is useful for all kinds of scientific image editing where you need a trail of your operations.

    1. 2

      Now that everyone has 32GB in their pocket, memory usage seems like a completely silly consideration.

      If image processing is happening concurrently, across multiple threads/processes on backend servers, constraining memory to something that will not bring your whole system down – is still an important consideration.