Call for Papers & Author Guidelines — Subspaces in the Era of Foundation Models (ACCV 2026)

Call for Papers

We invite original submissions on the theory, methods, and applications of subspace and manifold geometry in modern machine learning and computer vision:

Theory & Methods

Subspace learning & subspace geometry
Low-rank adaptation & parameter-efficient fine-tuning
Neural collapse & mechanistic interpretability
Optimization on Grassmann, Stiefel & Flag manifolds
Canonical correlation analysis & subspace clustering
Embedding methods & intrinsic dimensionality

Applications

Recognition & domain adaptation
Anomaly detection
Video analysis
3D reconstruction
Multimodal & vision–language alignment
Model compression & interpretability of foundation models

Accepted papers are peer-reviewed (double-blind) and will appear in the ACCV 2026 workshop proceedings, published by Springer in the LNCS series — a citable, archival publication.

Important dates

All deadlines are at 23:59 GMT.

Milestone	Date
Paper submission deadline	25 September 2026
Author notification	15 October 2026
Camera-ready deadline	27 October 2026
Workshop	14 December 2026 (AM)*

*Morning session, Grand Cube Osaka, Room 1001. Detailed clock times to be posted once finalized.

Submission

Papers must be prepared using the Springer LNCS template, up to 14 pages including all figures and tables, with unlimited additional pages for references only. All submissions must be in PDF format.

Format

Springer LNCS, up to 14 pages (references unlimited).

Review

Double-blind, at least two reviews per paper. No rebuttal this edition.

Platform

Via OpenReview.

LaTeX template: use the official ACCV 2026 LNCS template, available on the ACCV 2026 author guidelines page.

Policies

Aligned with ACCV 2026:

Submissions must be prepared for double-blind review: omit acknowledgements, institutional/lab names, grant numbers, and any links that reveal authorship; cite closely related own work in a way that preserves anonymity. Papers that reveal author identity, exceed the page limit, or violate formatting requirements may be desk-rejected without review.
Submissions must not contain hidden prompts, invisible text, or other concealed content intended to manipulate reviewers or automated review tools. LLMs must not be used to generate a submission, replace the authors' intellectual contribution, or fabricate/alter/obscure content; limited language or formatting assistance is acceptable with authors retaining full responsibility. External links that expand the submission's content, or that could bypass anonymity, page limits, or deadlines, are not allowed — all material for review must be included directly, in anonymized form.
Submissions must not be under review at another archival venue during the review period, nor substantially overlap with prior or concurrent published work. arXiv preprints are allowed.
Plagiarism detection may be applied.
Supplementary material must be anonymized and self-contained.
Submissions, reviews, and any author communications are confidential and must not be shared outside the review process.
Authors must maintain an up-to-date OpenReview profile and provide accurate conflict-of-interest information; incomplete or incorrect declarations may result in rejection without review.
For an accepted paper to appear in the proceedings, at least one author must register for the workshop and present the paper in person.
Authors should not publicly announce a submission as accepted before the official notification.