Title: ObjectNet
Description: ObjectNet is a large real-world test set for object recognition with control where object backgrounds, rotations, and imaging viewpoints are random. This work opens up new avenues for research in generalizable, robust, and more human-like computer vision and in creating datasets where results are predictive of real-world performance.
Keywords:
Mail addresses
objectnet@mit.edu
Domain: objectnet.dev
Links:
| https://objectnet.dev/index.html | |
| Download | https://objectnet.dev/download.html |
| Team | https://objectnet.dev/team.html |
| "How hard are computer vision datasets? Calibrating dataset difficulty to viewing time" | https://objectnet.dev/mvt/ |
| download | https://objectnet.dev/download.html |
| Paper | https://objectnet.dev/objectnet-a-large-scale-bias-controlled-dataset-for-pushing-the-limits-of-object-recognition-models.pdf |
| Competition | https://eval.ai/web/challenges/challenge-page/726/overview |
| PyTorch code | https://github.com/abarbu/objectnet-template-pytorch |
| TensorFlow code | https://github.com/abarbu/objectnet-template-tensorflow |
| Download | https://objectnet.dev/download.html |
| Spoken ObjectNet | https://github.com/iapalm/Spoken-ObjectNet |
| Image viewing times | https://objectnet.dev/flash/index.html |
Viewport: width=device-width, initial-scale=.4