Paper Detail

Modulate-and-Map: Crossmodal Feature Mapping with Cross-View Modulation for 3D Anomaly Detection

Alex Costanzino, Pierluigi Zama Ramirez, Giuseppe Lisanti, Luigi Di Stefano

arxiv Score 7.8

Published 2026-04-02 · First seen 2026-04-04

General AI

Abstract

We present ModMap, a natively multiview and multimodal framework for 3D anomaly detection and segmentation. Unlike existing methods that process views independently, our method draws inspiration from the crossmodal feature mapping paradigm to learn to map features across both modalities and views, while explicitly modelling view-dependent relationships through feature-wise modulation. We introduce a cross-view training strategy that leverages all possible view combinations, enabling effective anomaly scoring through multiview ensembling and aggregation. To process high-resolution 3D data, we train and publicly release a foundational depth encoder tailored to industrial datasets. Experiments on SiM3D, a recent benchmark that introduces the first multiview and multimodal setup for 3D anomaly detection and segmentation, demonstrate that ModMap attains state-of-the-art performance by surpassing previous methods by wide margins.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
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{costanzino2026modulate,
  title = {Modulate-and-Map: Crossmodal Feature Mapping with Cross-View Modulation for 3D Anomaly Detection},
  author = {Alex Costanzino and Pierluigi Zama Ramirez and Giuseppe Lisanti and Luigi Di Stefano},
  year = {2026},
  abstract = {We present ModMap, a natively multiview and multimodal framework for 3D anomaly detection and segmentation. Unlike existing methods that process views independently, our method draws inspiration from the crossmodal feature mapping paradigm to learn to map features across both modalities and views, while explicitly modelling view-dependent relationships through feature-wise modulation. We introduce a cross-view training strategy that leverages all possible view combinations, enabling effective an},
  url = {https://arxiv.org/abs/2604.02328},
  keywords = {cs.CV},
  eprint = {2604.02328},
  archiveprefix = {arXiv},
}

Metadata

{}