Paper Detail

LACToS: Local Anomaly Detection via Conditional Typicality of Superpixels

Jihoon Hong, Alireza Kheirandish, Rohan Sanda, Sara Fridovich-Keil

arxiv Score 11.3

Published 2026-10-22 · First seen 2026-10-23

General AI

Abstract

Local anomaly detection is the task of identifying spatially confined anomalous or out-of-distribution (OOD) regions within an otherwise in-distribution (ID) image. Most existing methods are designed for a specific context, with a strong focus on industrial defect detection, where they achieve excellent performance on well-established benchmarks. However, we observe that these methods can exhibit significant performance drops in other contexts, especially when the inlier distribution is multimodal. To account for this inlier complexity, we propose LACToS, a principled local OOD detection method based on statistical typicality and made tractable through two innovations: (i) grouping pixels into superpixels and (ii) using a computationally efficient proxy to estimate typicality. We evaluate LACToS across three diverse applications spanning medical imaging, human faces, and X-ray security screening. LACToS performs more robustly than prior methods across these benchmarks. The code is released here: https://github.com/voilalab/LACToS.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{hong2026lactos,
  title = {LACToS: Local Anomaly Detection via Conditional Typicality of Superpixels},
  author = {Jihoon Hong and Alireza Kheirandish and Rohan Sanda and Sara Fridovich-Keil},
  year = {2026},
  abstract = {Local anomaly detection is the task of identifying spatially confined anomalous or out-of-distribution (OOD) regions within an otherwise in-distribution (ID) image. Most existing methods are designed for a specific context, with a strong focus on industrial defect detection, where they achieve excellent performance on well-established benchmarks. However, we observe that these methods can exhibit significant performance drops in other contexts, especially when the inlier distribution is multimod},
  url = {https://arxiv.org/abs/2610.31538},
  keywords = {cs.CV, cs.LG},
  eprint = {2610.31538},
  archiveprefix = {arXiv},
}

Metadata

{}