From malin.sandstrom at incf.org Mon Jan 4 09:58:34 2021 From: malin.sandstrom at incf.org (=?UTF-8?Q?Malin_Sandstr=C3=B6m?=) Date: Mon, 4 Jan 2021 15:58:34 +0100 Subject: [Neuroimaging] Call for community project proposals and mentors for INCF's participation in GSoC 2021 - expression of interest due January 15 Message-ID: Hello all, ------- Please forward this email to prospective mentors you might have in your network ---- INCF will be applying again to Google Summer of Code, for the 11th time running. Last year was a record year in the number and diversity of projects, we hope to do at least as well this year! A note on scope: we are looking for open source project ideas with some sort of *neuro connection* - fixes or extensions to big/broadly applicable or small/niche tools used by neuroscience researchers, efforts and initiatives in the broad field of computational neuroscience. If your project supports or implements *neuroscience standards or best practices*, we are especially happy (but all projects are treated equally, when it comes to slot assignments). This year, we are *also looking for mentors who come without their own projects *and are willing to take on student-proposed projects that fit their interests. As you may have seen earlier this year, Google has announced some changes to the program . Projects will now be SHORTER (10 weeks instead of 12) and entry thresholds for students will be diversified, which will allow more varied student backgrounds. In compensation, there will be more projects. This means, if you have a running project from previous years that you want to continue, you may need to *re-scope* its contents and aims a bit, maybe even split it in two (independent) parts. With a shorter time span and a more varied student body, where not all students will be as near to full-fledged computer scientists, we will need to put a bit more effort into *introductions, support *and *confidence-building*. INCF will help with this, please refer interested students to us at an early stage so we can onboard them into our 'potential students' community. Another important change: this time (and in years forward), we aim to *keep all prospective students in the INCF community*, regardless of if they are accepted to a project or not. They may well be candidates for next year, and also, since you will be putting actual effort into project discussions, we want to make sure it is useful. Also, the usual reminder: there is no guarantee that INCF is accepted this year until we hear from Google on *March 9*. Please be clear with this in your communication with students. The* timeline * is similar to the years before: Applications open near the *end of January (29th)*, they close near *mid February (19th)*. We aim to have a live project ideas list in mid January. Orgs are announced on *March 9*, and student applications are open *March 21 - April 19*. *Important*: last year we were near peak capacity in the number of project ideas on our final list, with a flurry of last-minute activity. We therefore want interested mentors to contact us (gsoc at incf.org or malin at incf.org) *early* with a declaration of interest and a short summary of the intended project, latest *January 15*. If you want to mentor but don't have a project idea, you are still welcome to join our list of potential mentors for ideas that originate with students. Actual project descriptions should ideally come by *January 22*, a week before we officially go live with the project list on *January 29*, but we will take on late requests from *previous mentors only* (including mentors of previous non-accepted projects) on a case-by-case basis if they are well described and properly scoped. Project template: Each submission should come *by email* and contain: 1) An *informative title* with relevant keywords (good-to-have: name of tool/project, name of programming language) 2) A *project description* including a brief general intro, motivation, aims, scope and skills/skill level needed. Please indicate 2-5 tech keywords (Python, C++, Java, SQL, REST, CUDA, ...) 3) Name of *lead mentor* + a *named co-mentor or backup mentor* who would be able and willing to back the main/contact mentor up in mentoring in case something unforeseen happens (can be more than one, as long as you are internally clear on who does what) 4) Any planned *longer absences* during the lead up/student interaction period (*February 1 - April 19*, especially the last week) or the project period (May 17 - *August 30*), and the plan for covering them. 5)* Format*: your choice - txt, doc, docx, pdf or a link to a webpage with the full description(s). Anything that can be copied and pasted. NB: Project ideas will be posted on *Neurostars.org* after the ideas list goes live, and prospective mentors are expected to *join the forum and be available* for questions from students during the run up to the announcement (late Jan - early March) and the student interactions period (March-April). This can be handled by email, mostly, if you prefer, once you have set your NeuroStars account up. Questions and project submissions: please email *malin at incf.org * AND cc *gsoc at incf.org *. Happy new year of coding! /Malin -- Malin Sandstr?m, PhD Community Engagement Officer malin.sandstrom at incf.org International Neuroinformatics Coordinating Facility Karolinska Institutet Nobels v?g 15 A SE-171 77 Stockholm Sweden http://www.incf.org -------------- next part -------------- An HTML attachment was scrubbed... URL: From carinaforster0611 at gmail.com Mon Jan 18 11:36:14 2021 From: carinaforster0611 at gmail.com (Carina Forster) Date: Mon, 18 Jan 2021 17:36:14 +0100 Subject: [Neuroimaging] nipy.algorithms.registration.histogram_registration fails for registering CT image to MRI Message-ID: Hello neuroimaging community, I want to align a CT image with a T1 MPRAGE to eventually mark ECoG electrodes on the T1. I aligned the T1 to acpc and used freesurfer for surface reconstruction. My next step would be to align the CT image with FSL output orig.mgz. My CT has the follwing dimensions: 512 x 512 x 734, image orientation: LAS My MRI: 256x256x256, image orientation: RAS It worked for a few images but not for all. nipy tells me the coregistration worked, however I noticed that the normalized mutual information is 0.2 for the images that are correctly aligned and about 2.0 for the ones where coregistration didn't work. My understanding is that NMI shoudl be between 0 and 1, so I think there is something wrong. If anyone has any idea what might go wrong, I am forever grateful. Have a great monday, Carina -------------- next part -------------- An HTML attachment was scrubbed... URL: