The four bars show ImageNet top-1 accuracy for a high-performance deep learning Resnet50 model for normal ImageNet test images and for data sets with degraded images. Human performance remains high, above 90%, even with degraded images, as shown by the upper shaded area.
This site uses cookies. By continuing to use our website, you are agreeing to our privacy policy. No content on this site may be used to train artificial intelligence systems without permission in writing from the MIT Press.