Latentic LabsPublished by Latentic Labs with scime atlasscime
The scime Dispatch

ECCV 2026 edition · every accepted paper, counted and compared

ECCV 2026 · 2,834 accepted papersNovember 2026Free · atlas.scime.ai/newspaper

← All editions

Multimodal LLMs: 3.4% of ECCV 2024, 11.8% this year

ECCV grew 19% to 2,834 accepted papers. Multimodal LLMs grew faster still, from 80 to 334.

Multimodal LLMs (Vision-language and multimodal) went from 80 papers at ECCV 2024 to 334 this year, from 3.4% to 11.8% of the conference. Inside it, Spatial reasoning in VLMs grew fastest (4 → 64 papers). Among them: ViewSpatial-Bench: Evaluating Multi-perspective Spatial Localization in Vision-Language Models and Video-Holmes: Can MLLM Think like Holmes for Complex Video Reasoning?.

Also gaining: Video generation (3.8% → 9.1%) and Robot action and navigation (0.6% → 3.3%). Losing ground: Text-to-image diffusion (6.1% → 3.8%) and Open-set and few-shot detection (4.4% → 2.1%).

Qwen is the model family ECCV 2026 papers build on most: 456 of 2,806 papers with result tables (16%), ahead of ViT (8%) and ResNet (5%). The benchmark named most is ImageNet, then COCO.

Not everything helped. Of the 35,414 changes the papers measured, 24% made no clear difference or made things worse, as the papers report them. One of 24 on page D:

A paper says it overturns

2,810 papers name an assumption their results go against; 20 of them on page E.

Accepted ECCV papers per year in scime atlas.

AI-assisted. Topics come from the atlas clustering; what each paper tested and measured was extracted by AI from its full text. Nothing here was checked by hand and some lines are wrong; open the paper before you cite it.

ECCV 2026 in numbers

Accepted papers2,834
ECCV 20242,387
Read in full by AI2,812
Findings recorded (change → effect)35,414
Reported as helping73%

Gained most, 2024 → 2026

Multimodal LLMs3.4 → 11.8%
Video generation3.8 → 9.1%
Robot action and navigation0.6 → 3.3%
Video-language understanding2.3 → 4.3%
Gaussian splatting4.0 → 6.0%

Lost most

Text-to-image diffusion6.1 → 3.8%
Open-set and few-shot detection4.4 → 2.1%
NeRF and radiance fields2.4 → 0.2%
Video segmentation and tracking4.3 → 2.6%

Largest areas, 2026

3D vision21.9%
Vision-language and multimodal20.6%
Video16.9%
Image generation14.0%

Try it on your idea

Describe your own idea in scime atlas to see the closest of 137,217 papers and what similar changes did in them.

Writing it up? scime draft is a free LaTeX editor with AI.

Free for signed-in readers

Read the rest of this edition

Five more pages: where the 2,834 papers are, the topics that grew and shrank, what 2,809 papers measured and when it did not help, what the papers say they overturn, and the benchmarks, baselines and models everyone uses. Sign in with a free scime account to keep reading.

  • B Where the papers are
  • C What grew, what shrank
  • D What the papers measured
  • E What the papers overturn
  • F Benchmarks, baselines and models

Sign in free

Already signed in? Reload this page. Accounts are free; we only use your email to sign you in and, if you agree, to tell you about new editions.

G

About the data

how this edition was made

Papers. The 2,834 accepted ECCV 2026 papers and 2,387 of 2024 in scime atlas, grouped with all 137,217 atlas papers (ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV, ECCV, TPAMI, IJCV and JMLR, 2016–2026) into 20 areas and their topics by meaning (an open-source embedding model and clustering; names written with AI).

Full texts. For 2,812 of the 2026 papers an AI model read the full text and recorded the result tables (benchmark, method, base model, number) and each change the paper tested, on what, the metric and whether it helped, plus what the paper says it challenges. These were not checked by hand. Each paper links to its page on scime atlas, which links to the paper itself.

Counting. Shares count each paper once. Citation counts are from OpenAlex and are still low for 2026 papers; papers are ordered by citations, then by how many findings were recorded.

Caveat. Published papers report successes more often than failures. Corrections: contact@latenticlabs.com. Not affiliated with or endorsed by ECCV or any of the conferences.

Get the next scime Dispatch and a monthly field update

One email when a new edition is out (ICLR, CVPR, ICML, NeurIPS, ACL and special editions), plus a short monthly note on what moved in machine learning research.

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy

← All editions · © 2026 Latentic Labs