Images

ObjectNet

Introduced by Barbu et al. in ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models

ObjectNet is a test set of images collected directly using crowd-sourcing. ObjectNet is unique as the objects are captured at unusual poses in cluttered, natural scenes, which can severely degrade recognition performance. There are 50,000 images in the test set which controls for rotation, background and viewpoint. There are 313 object classes with 113 overlapping ImageNet.

Source: On Robustness and Transferability of Convolutional Neural Networks

Homepage

Benchmarks

Add a new result Link an existing benchmark

Task	Dataset Variant	Best Model
Image Classification	ObjectNet	CoCa
Zero-Shot Transfer Image Classification	ObjectNet	LiT-22B
Image Classification	ObjectNet (Bounding Box)	BiT-L
Unsupervised Image Classification	ObjectNet	InfoMin ResNeXt-152 + SK