# ObjectNet Dataset: Reanalysis and Correction

arXiv: https://arxiv.org/abs/2004.02042 · HTML: https://mlsanity.com/paper/2004.02042/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Apr 2020.

## Built on by (most cited first, 5 of 5)

- [CrossNorm and SelfNorm for Generalization under Distribution Shifts](https://mlsanity.com/paper/2102.02811/) (2021)
- [Does Data Repair Lead to Fair Models? Curating Contextually Fair Data To Reduce Model Bias](https://mlsanity.com/paper/2110.10389/) (2021)
- [Contemplating real-world object classification](https://mlsanity.com/paper/2103.05137/) (2021)
- [MMLU-SR: A Benchmark for Stress-Testing Reasoning Capability of Large Language Models](https://mlsanity.com/paper/2406.15468/) (2024)
- [Reconstructing Content with Collaborative Attention for Universal Multimodal Representation Learning](https://mlsanity.com/paper/2603.01471/) (2026)

## Abstract (arXiv)

Recently, Barbu et al introduced a dataset called ObjectNet which includes objects in daily life situations. They showed a dramatic performance drop of the state of the art object recognition models on this dataset. Due to the importance and implications of their results regarding generalization ability of deep models, we take a second look at their findings. We highlight a major problem with their work which is applying object recognizers to the scenes containing multiple objects rather than isolated objects. The latter results in around 20-30% performance gain using our code. Compared with the results reported in the ObjectNet paper, we observe that around 10-15 % of the performance loss can be recovered, without any test time data augmentation. In accordance with Barbu et al.'s conclusions, however, we also conclude that deep models suffer drastically on this dataset. Thus, we believe that ObjectNet remains a challenging dataset for testing the generalization power of models beyond datasets on which they have been trained.