A 2017 paper using bob.bio

[Manuel Günther]

Just a note, in case you haven't seen this yet:

Did you ever want to know, what are the baseline results of all of the algorithms in bob.bio on most of our standard face evaluation databases? It seems that there have been some PhD students from India (as my Indian colleague has verified) who took the facereclib and some of our database interfaces, ran an evaluation with the default parameters and published a paper called:

"Performance Comparison of Some Face Recognition Algorithms on Multi-covariate Facial Databases" https://ieeexplore.ieee.org/document/8313741

in a tier-two or even tier-three conference: International Conference on Image Information Processing (ICIIP)

Why didn't we have this idea? ;-)

Manuel

[Tiago Freitas Pereira]

Hey Manuel,

Thanks for sharing this, that's awesome.
However, this is kinda similar with what you have done in http://publications.idiap.ch/downloads/papers/2016/Gunther_SPRINGER_2016.pdf

Actually, the book chapter is more complete, don't you think?

Cheers

--
-- You received this message because you are subscribed to the Google Groups bob-devel group. To post to this group, send email to bob-...@googlegroups.com. To unsubscribe from this group, send email to bob-devel+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/bob-devel or directly the project website at http://idiap.github.com/bob/
---
You received this message because you are subscribed to the Google Groups "bob-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bob-devel+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Tiago

[Manuel Günther]

Tiago,

Oh, yes, the book chapter is definitely more complete. It was also much more time consuming to design and run the experiments. 

The mentioned paper did a very low-effort approach here, by just taking the default parameters and report the default evaluations. I never thought that this might be sufficient to publish a paper, that's why I didn't try it. Maybe we should have ;-)

Manuel