search

visionjo / facerec-bias-bfw

Source code and notebooks to reproduce experiments and benchmarks on Bias Faces in the Wild (BFW).
BSD 3-Clause "New" or "Revised" License
45 stars 9 forks source link

Create a wrapper function to unify pipeline that produces the 3 figures (detailed below) from embedding data #9

Closed suchanv closed 4 years ago

suchanv commented 4 years ago

3 Figures based on the paper "Face Recognition: Too Bias, or Not Too Bias" are 1) DET curves: FPR versus FNR by moving threshold 2) Score distributions for genuine and imposter using violin plots 3) Confusion matrix for Rank 1 and any Rank.

visionjo commented 4 years ago

This is both ambiguous and broad in scope.

For starters, "plots we want" is very unclear from where I am sitting (I mean, I am sure I can guess, but, still, I might forget tomorrow-- say i work on a different project all night and return to this in morning, I would have to regroup, making this concept less effect)

Secondly, this is a large task-- too large without subtasks

An issue per task. A PR per issue. Then, issues like this can be made that reference the other issues for which make it up (i.e., end-to-end script to generate plots from ISSUE A, B, C).

I only mention this so you learn. It works for me however you do it, but if you want me to set deadlines for myself then it must be refined. I am not confident estimating time for this task, as it is quite unclear what it entails (open to perspective)

suchanv commented 4 years ago

Got it. Sorry for the ambiguity. However, I don't think it is broad at all if we have all the functions in a nice format. Let me try to make it more clear and you can comment.

This can be also on me if you do not have time.