ECCV 2026 workshop on deep learning era SfM set for Sept 9 with 3 speakers and 8 papers
ducha_aiki · x · 2026-09-07
ECCV 2026 will host the "Structure-from-Motion in the Age of Deep Learning" workshop on the morning of September 9 in Malmö, Sweden, featuring 3 speakers and 8 papers already available on OpenReview.
The workshop examines how modern SfM pipelines increasingly embed deep learning in feature matching, correspondence filtering, depth/pose initialization, and even end-to-end reconstruction, while the core remains geometric: relative pose estimation, calibration, triangulation, and bundle adjustment built on minimal solvers and robust estimation. Learning-based methods still struggle with cross-domain generalization, adverse imaging conditions, and uncertainty modeling, whereas classical pipelines fail on low-texture scenes, extreme viewpoints, dynamic content, and repetitive structures.
The framing: learning and geometry form an evolving design space to be integrated coherently, not a replacement story.
More from Research
- TFO: MBZUAI Brings Speech-Centric Omni Understanding to Frozen VLMs Without Retraining — MBZUAI · 2026-09-07
- NAVER AI Study Finds Distinct LLM Reasoning Operations Are Geometrically Separable in Hidden States — naver-ai · 2026-09-07
- SRMA: A Game-Theoretic Bilevel Approach Brings Convergence Guarantees to Multi-Agent LLM Systems — Yihang Chen · 2026-09-07
- Training mixtures are now all synthetic: small-model training is really distillation — RexDouglass · 2026-09-07
- DeepMind's Astra builds its own metamaterials lab, finds design with 2.1x strength ratio — ProfBuehlerMIT · 2026-09-07
- Tencent open-sources EVIE visual document retrieval models, hitting 66.75 nDCG@10 on ViDoRe V3 — jacek2023 · 2026-09-07