1. About
  2. Features
  3. Explore

I have some previously published work in the computer vision domain, and would like to upload them to arXiv to provide greater accessibility.

How can I go about finding which licence statement to select for the particular conferences my work was published in (CVPR & ECCV)?

1 Answer 1

Take a look at the copyright agreement you signed when you published the paper. It usually states under which conditions uploading the paper to other repositories is acceptable. Usually, you are only allowed to upload the "authors version" (or some variant thereof), which usually needs to contain a specific phrase like (in the case of the ACM):

© ACM, YYYY. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in PUBLICATION, {VOL#, ISS#, (DATE)} http://doi.acm.org/10.1145/nnnnnn.nnnnnn"

There is also a web service that makes this a bit easier to figure out: SHERPA/RoMEO. There you can put in the publisher / journal, and it will give you a human-readable summary of their policies (for example, here is the page for the ACM). But note that publisher policies change, and the binding version is the one you signed when you published the paper, so the data from SHERPA/RoMEO may be incorrect.

Note that, in my experience, almost all publishers allow uploading to an "institutional repository", so even if ArXiv may be disallowed, you can still make your work accessible via your institutional repository (if your institution has one) or your homepage. But, again, check the specific policies that apply to you on that.

Finally: I am not a lawyer, but as I understand it, since you no longer own the copyright on the paper, you cannot put it under a creative commons license (which is an option on ArXiv). So, you'll have to take the minimal license they allow, which only gives them permission to distribute your work (If someone else knows this better than I do, feel free to correct me in the comments or edit the answer).

Also, you can always just ask the publisher if you want to be extra sure, but in my experience, this may be a long process (I waited for over a week for a reply by the ACM, and am have now been waiting for over a month for a reply to a followup question).