3 ms·
imagenet is a truly terrible dataset. Any image recognition task that humans are not >99.9% on is not an image recognition task.
by ouid 9y ago
imagenet is a truly terrible dataset. Any image recognition task that humans are not >99.9% on is not an image recognition task.
- tanilama 9y agoA truly terrible dataset that leads to AlexNet/VGG/Neural Style Transfer/GoogleNet/ResNet, and facilitate research so much, is still a good dataset. Actually no, it is a phenomenal dataset. Not to mention, it is free and open to all. You can make assessment even if you are not an expert, but I would think you need more efforts to backup your claim a little bit to convince people and avoid downvoting.
- PeterisP 9y ago99.9% is an extremely high bar. People aren't machines, people make errors. If you asked them to tell whether a square is black or white, you would get less than 99.9%.
- ouid 9y ago99.9% is a very low bar for image recognition. if one out of every thousand times you saw a car you misclassified it as a chair, you would be dead.
- mturmon 9y agoIt sounds like you may not have experience in this domain? You're benchmarking an active, time-based task (a person "seeing a car" - implicitly allowing eye saccade, multiple images, context, depth) versus a static task ("classify this single image based on no context"). They are not comparable.
- ouid 9y agohave you actually tried the imagenet task yourself? The reason imagenet is popular is because it has allowed google to claim that they can beat human performance on image recognition by artifically lowering human performance on image recognition.
- nullc 9y agoYou mean you can't personally identify all 120 dog breeds in imagenet?
- eanzenberg 9y agoYou should know that for a lot of menial tasks machines are more accurate because there is no boredom, and boredom breeds mistakes.