A survey on bias in visual datasets

HIGHLIGHTS

SUMMARY

Computer vision (CV), the primary domain that enables computers to gain high-level understanding from visual data, is heavily dominated nowadays by deep learning (DL) methods that allowed for outstanding performance in tasks like object detection, image classification and image segmentation. For instance, faVisual Data cial recognition technologies, gender classification algo- In Caltech101 pictures labelled as cars are usually taken from the side, while ImageNet contains more racing cars. For instance, Terhörst et_al found some evidence that the face recognition dataset Labelled Faces in the Wild (LFW, for the original release and_[Kumar . . .

 



If you want to have access to all the content you need to log in!

Thanks :)

Username or Email

Password

Remember me

Lost your password?

If you don't have an account, you can create one here.