AnoDDPM: Anomaly Detection with Denoising Diffusion Probabilistic Models using Simplex Noise

By Julian Wyatt, Adam Leach, Sebastian M. Schmon, Chris G. Willcocks. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2022

484 citations recorded by the source corpus.

Generative models have been shown to provide a powerful mechanism for anomaly detection by learning to model healthy or normal reference data which can subsequently be used as a baseline for scoring anomalies. In this work we consider denoising diffusion probabilistic models (DDPMs) for unsupervised anomaly detection. DDPMs have superior mode coverage over generative adversarial networks (GANs) and higher sample quality than variational autoencoders (VAEs). However, this comes at the expense of poor scalability and increased sampling times due to the long Markov chain sequences required. We observe that within reconstruction-based anomaly detection a full-length Markov chain diffusion is not required. This leads us to develop a novel partial diffusion anomaly detection strategy that scales to high-resolution imagery, named AnoDDPM. A secondary problem is that Gaussian diffusion fails to capture larger anomalies; therefore we develop a multi-scale simplex noise diffusion process that gives control over the target anomaly size. AnoDDPM with simplex noise is shown to significantly outperform both f-AnoGAN and Gaussian diffusion for the tumorous dataset of 22 T1-weighted MRI scans (CCBS Edinburgh) qualitatively and quantitatively (improvement of +25.5% Sørensen–Dice coefficient, +17.6% IoU and +7.4% AUC).

Subfields: Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/20f6fce7726e7b3ab4ca45ef40d92b79f093f825

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Ming-Hsuan Yang
2. Luc Van Gool
3. Mubarak Shah
4. Radu Tudor Ionescu
5. Fahad Shahbaz Khan
6. Trevor Darrell
7. Xiaogang Wang
8. Dacheng Tao

In Radiology, Nuclear Medicine and Imaging

1. Ehsan Khodapanah Aghdam
2. Moein Heidari
3. Ming-Hsuan Yang
4. Dorit Merhof
5. Reza Azad
6. Ilker Hacihaliloglu
7. Amirhossein Kazerouni
8. Rui Gao

In Statistics and Probability

1. Bin Cui
2. Mubarak Shah
3. Yue Zhao
4. Ming-Hsuan Yang
5. Wentao Zhang
6. Shenda Hong
7. Ling Yang
8. Zhilong Zhang

Citations in the corpus

• Diffusion Models in Vision: A Survey (2022, background): “[46] train a DDPM on healthy medical images.”
• Revisiting Reverse Distillation for Anomaly Detection (2023, methodology): “Comparison between Simplex noise [37] and Gaussian noise in simulating the pseudo anomalous regions in the image.”
• Diffusion models in medical imaging: A comprehensive survey (2022, methodology): “AnoDDPM [70] substantiates that generalization to other types of noise distributions can enhance task-specific quality.”
• AnomalyGPT: Detecting Industrial Anomalies using Large Vision-Language Models (2023, methodology): “RIAD [33], SCADN [30], InTra [21] and AnoDDPM [28] employ different reconstruction network architectures, ranging from autoencoder and Generative Adversarial Network (GAN) to Transformer and diffusion model.”
• Diffusion Models: A Comprehensive Survey of Methods and Applications (2022, methodology): “Generative models have been shown to own a powerful mechanism for anomaly detection [70, 87, 252], modeling normal or healthy reference data.”