Diffusion Imaging in Python | Hacker News Reader

Diffusion Imaging in Python

nipy.org

Could this be used for images other than people's brains?

I looked at this neuroscience course, and overwhelmingly it was a course on Fourier transforms and Radon transforms

http://www.cogsci.ucsd.edu/~sereno/276/notes/recon2.gif

Momentarily, I was asking myself why such a difficult math course is being given in the Psychology department

certainly, diffusion imaging is common in analysis of many physiologies including breast and prostate. It forms one of the sequences in multiparametric MRI which has had some recent success in improving the diagnosis of prostate cancer [1].

I'm also aware of at least one upcoming trial of full (or mostly full) body diffusion imaging to look for sites of metastasis. The trial is still in planning so unfortunately no link to share for that one.

[1]: http://thelancet.com/journals/lancet/article/PIIS0140-6736(1...

import antigravity [1]

I in Python, Given the fact Python is a dynamically typed language, with all the need to write deterministic tests, I am always amazed to see such packages come up. The eco-system is just incredible in this regard. Thumbs up!

[1] https://xkcd.com/353/

i am always conflicted when using python, while it is unusually productive to hack up something in python i always keep on worrying about performance. I wish i could code in python and get performance of say golang.

This is a recurring theme in the field.

You could consider Julia, especially now that has received some serious funding. But its problem is that it's still not v1.0, there may be serious changes before that milestone.

I don't think that Julia's funding is a good reason to pick it up, maybe even the opposit of that.

Julia is really going to explode once it hits 1.0 status. I feel like that's the main thing holding people back from contributing to the library ecosystem.

Are there any good examples of Julia doing serious work that can't be done in Python?

With libraries like Numba in Python I just can't see the reason for Julia (aside from the neat type system but that's not enough to switch imo)

Take a look at Celeste [1], which ran on the Cori supercomputer. At that level you really need high performance not just for numerics but also for IO and heavily custom data structures, which was previously only possible in Fortran and C++.

[1]: https://juliacomputing.com/case-studies/intel-astro.html

Most of the time it's not an issue. Especially in numerical work it is though, and libs like numba have your back.

Tests are useful even in typed langs. Though Python is dynamic, the sort of code you want in production is statically analysable and tools like flake8 and mypy can find many of the issues that a a compiler would find in a static strong language.

Reply on news.ycombinator.com