# A Study and Comparison of Human and Deep Learning Recognition Performance Under Visual Distortions

arXiv: https://arxiv.org/abs/1705.02498 · HTML: https://mlsanity.com/paper/1705.02498/

## Key facts

- Cited by 108 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 May 2017.

## Built on by (most cited first, 30 of 108)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Partial success in closing the gap between human and machine vision](https://mlsanity.com/paper/2106.07411/) (2021)
- [3D Common Corruptions and Data Augmentation](https://mlsanity.com/paper/2203.01441/) (2022)
- [On Interaction Between Augmentations and Corruptions in Natural Corruption Robustness](https://mlsanity.com/paper/2102.11273/) (2021)
- [Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision](https://mlsanity.com/paper/2202.08360/) (2022)
- [Human alignment of neural network representations](https://mlsanity.com/paper/2211.01201/) (2022)
- [On the effectiveness of adversarial training against common corruptions](https://mlsanity.com/paper/2103.02325/) (2021)
- [On the Design Fundamentals of Diffusion Models: A Survey](https://mlsanity.com/paper/2306.04542/) (2023)
- [A Comprehensive Survey of Foundation Models in Medicine](https://mlsanity.com/paper/2406.10729/) (2024)
- [If your data distribution shifts, use self-learning](https://mlsanity.com/paper/2104.12928/) (2021)
- [Enhancing Diffusion-Based Image Synthesis with Robust Classifier Guidance](https://mlsanity.com/paper/2208.08664/) (2022)
- [A Review of Adversarial Attack and Defense for Classification Methods](https://mlsanity.com/paper/2111.09961/) (2021)
- [MMG-Ego4D: Multi-Modal Generalization in Egocentric Action Recognition](https://mlsanity.com/paper/2305.07214/) (2023)
- [Robustness via Cross-Domain Ensembles](https://mlsanity.com/paper/2103.10919/) (2021)
- [Self-attention in Vision Transformers Performs Perceptual Grouping, Not Attention](https://mlsanity.com/paper/2303.01542/) (2023)
- [Do Perceptually Aligned Gradients Imply Adversarial Robustness?](https://mlsanity.com/paper/2207.11378/) (2022)
- [Neural Architecture Design and Robustness: A Dataset](https://mlsanity.com/paper/2306.06712/) (2023)
- [LLM Whisperer: An Inconspicuous Attack to Bias LLM Responses](https://mlsanity.com/paper/2406.04755/) (2024)
- [Progress and limitations of deep networks to recognize objects in unusual poses](https://mlsanity.com/paper/2207.08034/) (2022)
- [Comparing object recognition in humans and deep convolutional neural networks -- An eye tracking study](https://mlsanity.com/paper/2108.00107/) (2021)
- [Anti-aliasing Deep Image Classifiers using Novel Depth Adaptive Blurring and Activation Function](https://mlsanity.com/paper/2110.00899/) (2021)
- [DeepAdversaries: Examining the Robustness of Deep Learning Models for Galaxy Morphology Classification](https://mlsanity.com/paper/2112.14299/) (2021)
- [Rapid Network Adaptation: Learning to Adapt Neural Networks Using Test-Time Feedback](https://mlsanity.com/paper/2309.15762/) (2023)
- [Rethinking Natural Adversarial Examples for Classification Models](https://mlsanity.com/paper/2102.11731/) (2021)
- [Benchmarks for Corruption Invariant Person Re-identification](https://mlsanity.com/paper/2111.00880/) (2021)
- [Imperceptible Protection against Style Imitation from Diffusion Models](https://mlsanity.com/paper/2403.19254/) (2024)
- [MomentumSMoE: Integrating Momentum into Sparse Mixture of Experts](https://mlsanity.com/paper/2410.14574/) (2024)
- [Closer Look at the Uncertainty Estimation in Semantic Segmentation under Distributional Shift](https://mlsanity.com/paper/2106.00076/) (2021)
- [Failure Modes of Domain Generalization Algorithms](https://mlsanity.com/paper/2111.13733/) (2021)
- [Test-time Batch Normalization](https://mlsanity.com/paper/2205.10210/) (2022)

## Abstract (arXiv)

Deep neural networks (DNNs) achieve excellent performance on standard classification tasks. However, under image quality distortions such as blur and noise, classification accuracy becomes poor. In this work, we compare the performance of DNNs with human subjects on distorted images. We show that, although DNNs perform better than or on par with humans on good quality images, DNN performance is still much lower than human performance on distorted images. We additionally find that there is little correlation in errors between DNNs and human subjects. This could be an indication that the internal representation of images are different between DNNs and the human visual system. These comparisons with human performance could be used to guide future development of more robust DNNs.