Full papers
Original submissions on theory, methods, and applications of subspace and manifold geometry in modern machine learning and computer vision; peer-reviewed and published in Springer LNCS proceedings.
AI/CORE Unranked
Subspaces ACCV26 2026
Subspaces in the Era of Foundation Models (ACCV 2026 Workshop)
The Subspaces 2026 workshop, part of ACCV 2026 in Osaka, Japan, explores the connection between classical subspace and manifold geometry and the low-dimensional structures within modern foundation models in vision and language. It aims to bridge theoretical advances in subspace learning with practical applications in machine learning and computer vision.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Research areas in scope
01
Theory & Methods
Subspace learning & subspace geometryLow-rank adaptation & parameter-efficient fine-tuningNeural collapse & mechanistic interpretabilityOptimization on Grassmann, Stiefel & Flag manifoldsCanonical correlation analysis & subspace clusteringEmbedding methods & intrinsic dimensionality
02
Applications
Recognition & domain adaptationAnomaly detectionVideo analysis3D reconstructionMultimodal & vision–language alignmentModel compression & interpretability of foundation models
Policies worth checking twice
- Submissions must be prepared for double-blind review: omit acknowledgements, institutional/lab names, grant numbers, and any links that reveal authorship; cite own work anonymously.
- Papers must not exceed 14 pages (including figures and tables), with unlimited references; must use the Springer LNCS template and be in PDF format.
- No rebuttal process this edition.
- Submissions must not contain hidden prompts, invisible text, or concealed content intended to manipulate reviewers or automated tools.
- LLMs must not be used to generate submissions, replace authors' intellectual contribution, or fabricate/alter/obscure content; limited language or formatting assistance is acceptable with full author responsibility.
- External links that expand content, bypass anonymity, page limits, or deadlines are not allowed; all review material must be included directly in the submission.
- 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.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026