Newest papers

LACToS: Local Anomaly Detection via Conditional Typicality of Superpixels

Jihoon Hong, Alireza Kheirandish, Rohan Sanda, Sara Fridovich-Keil · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31538 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

La detección de anomalías locales es la tarea de identificar regiones de anomalías confinadas espacialmente o fuera de distribución (OOD) dentro de una imagen de otra manera en distribución (ID). La mayoría de los métodos existentes están diseñados para un contexto específico, con un fuerte enfoque en la detección de defectos industriales, donde obtienen un excelente rendimiento en benchmarks bien establecidos. Sin embargo, observamos que estos métodos pueden mostrar caídas significativas de rendimiento en otros contextos, especialmente cuando la distribución interior es multimodal. Para tener en cuenta esta complejidad interior, proponemos LACToS, un método de detección local de OOD basado en principios basados en la tipicidad estadística y hecho tratable a través de dos innovaciones: (i) agrupar píxeles en superpixeles y (ii) usar un pro

Machine translation (Meta M2M100); the English original follows.

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.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Anomaly detection, Artificial intelligence, Computer science, Focus (optics), Pattern recognition (psychology), Pixel

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31538, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange