Identifying trannies with AI -

hundredpercent

kiwifarms.net
If you are using Facenet then it probably has a way to consider multiple faces in an image. Just pick the face with the highest likelihood. Regardless, unless the data is massively tainted then a classifier should still work despite some false positives. Ditto for Asians, unless they are 90% of the dataset, the network will find how to deal with them. If you have labels for ethinicity, it might help to add an additional race class output to the network that can boost performence.
Highest likelihood of what? They're both faces, but the images are labeled by class. If there's an image marked "trap" and it contains one trap and one man, then it's 50/50 which one gets marked trap and which one discarded.
Transfer learning should already work with a small dataset, especially when you have thousands of images. Dimensionality reduction won't help in any way. What is the input to your network? Maybe try to use something YOLO to cut a square around every person and input that to your network, to reduce background noise.
I am using sklearn's MLPClassifier, inputs are FaceNet's outputs. The accuracy is slightly better than SVM. Adding noise to the images won't change anything.
 

Smaug's Smokey Hole

no corona
kiwifarms.net
Wasn't the problem with previous facial recognition scandals that it was trained on men and women and correctly pointed out the men in women's clothing? What you are creating now is something to differentiate between weird pervs trying to look like women taking it up the ass and unnatural plastic looking women taking it up the ass. Poor Skynet...

Random picture from your dataset (why is it all porn?) "Passing" troon or not?
thumb_279071_crop.jpg

thumb_279071.jpg

Bonus cursed image
thumb_296601.jpg
 

hundredpercent

kiwifarms.net
What you are creating now is something to differentiate between weird pervs trying to look like women taking it up the ass and unnatural plastic looking women taking it up the ass.
Good point. Maybe I should use a better dataset for the women.
why is it all porn
Where else to find 7000+ images of trannies?

I'm at 80% accuracy or so and it seems like I might be able to get it to 90% with some finagling, but I've just realized it's pointless. Even if it's 95% accurate, only 0.3% of the population are trannies, which means the probability of it actually being a tranny given a positive match is P(A|B) = P(B|A)*P(A)/P(B) = .95*.003 / (.95*.003 + .05*.997) = 0.05 = 5%.

If anyone wants the dataset and code I'll post it, but it seems pretty pointless otherwise.
 
Tags
None