15 ms·
https://www.imageidentify.com/ https://www.imageidentify.com/ correctly identified the images as "a small sofa". I think rotating the image is questionable, sin
by xyproto 11y ago
https://www.imageidentify.com/ https://www.imageidentify.com/ correctly identified the images as "a small sofa". I think rotating the image is questionable, since there could be an algorithm for first orienting the image correctly based on light and shadow and then the image recognition could be run.
https://www.imageidentify.com/result/1ixb9603m9ix1 https://www.imageidentify.com/result/1ixb9603m9ix1
- FranOntanaya 11y agoIt could be that the dataset is so large there's very close matches to any common given picture, and the algorithm is actually awful at picking up details. Giving it an image that we know has all the relevant details of a sofa, but it likely won't have close matches in the dataset, can give us an idea of how clever it is.
- quasiresearcher 11y agoBut those algorithms would be very limited in usefulness. Systems such as imageidentify.com should be at least trying an ensemble of algorithms, many of which I suppose should be invariant under translation and rotation. Edit: There's a comment about invariance in this thread [1] and apparently CNNs are not invariant under rotation. [1] https://news.ycombinator.com/item?id=9750133 https://news.ycombinator.com/item?id=9750133
- sbodenstein 11y agoWe did not train ImageIdentify to be invariant under arbitrarily large rotations. This is fairly easy to do: show the network couches rotated at all angles.