The first ICCV 2025 Workshop
Findings


Oct 20 8:45am-5pm @ ICCV 2025 in Honolulu, HI, USA

Venue: Honolulu Convention Center, Room 310 | There will be no virtual component


Overview

This workshop introduces the concept of a findings-style track to the computer vision community. The number of submissions to computer vision conferences has grown considerably over the years, posing scaling challenges to organizers.


Number of Submissions at CVPR 2006-2024 (taken from Opening Remarks at CVPR'24)


Findings tracks have been used in the NLP community since 2020 and can help with these challenges by reducing resubmissions. They accomplish by publishing technically sound work, but which may not meet the main conference's threshold for novelty, impact, or excitement. There are many important results that the community should be made aware of, and this venue ensures an audience for these results without delays for further iterations of submissions or that might otherwise be lost if never published. This workshop provides a vehicle to discuss related experience and challenges with integrating a Findings track at computer vision conferences in an effort to advocate for a fully integrated Findings track as well as present Findings-quality papers to help highlight their potential impact and benefits to inform future conferences.

Call for Papers

We call for full papers (8 pages) to be presented at the 1st Workshop on the Findings of ICCV. This workshop aims to publish technically sound and well-executed papers, which may not meet the main conference's threshold for novelty or impact, similar to those accepted to the Findings track at NLP conferences since 2020. There are many important results that the community should be made aware of, and this venue ensures an audience for these results without delays for further iterations of re-submissions or that might otherwise be lost if never published. Thus, papers accepted by this workshop will contain solid, complete work on a broad spectrum of topics, unlike typical workshops focused on a particular topic and which may also accept works in progress. Accepted papers will be included in the conference proceedings (i.e., if accepted, it cannot be resubmitted elsewhere) with a presentation opportunity in the workshop.

To that end, submitted papers should be withdrawn or rejected from the main conference of ICCV’25 and will receive a new meta-review based on the Findings criteria. Authors are asked to submit their originally submitted paper, rebuttal, and any reviews or meta-reviews it received in the supplementary. Since accepted papers will be included in the conference proceedings, we will not accept papers that are under review at other venues. In addition, we will not consider works in progress that have not undergone at least one round of peer review.

We seek contributions on the same topics as the main conference (see the ICCV CfP for a list of topics). Submitted papers will be re-evaluated using the Findings criteria. These papers should 1) be technically sound, 2) teach something new, i.e. there is a finding, 3) have clear claims, method, and experiments (with minor revision), and 4) support all claims with evidence, with limitations clearly expressed (but experiments do not need to explore beyond the claims), with a presentation opportunity at the ICCV Findings Workshop.

Submission timeline


Paper submission: June 30th, 2025 23:59 US Hawaii Standard Time

Acceptance notification: ~July 9th, 2025


Submission materials (without any modification):

  • The paper version submitted (and subsequently rejected or withdrawn) to ICCV'25.
    • We will also consider papers submitted and rejected from ICML’25.
    • Note that papers should not be in submission to another venue.
  • In the supplementary, include:
    • The paper’s original supplementary material (if applicable).
    • A copy of the reviews, meta-review, and rebuttal.
OpenReview Submission Site

Keynote Speakers/Panelists

Lun-Wei Ku

Nihar Shah

Deqing Sun

Siyu Tang

Academia Sinica
Carnegie Mellon University
Google DeepMind
ETH Zürich



Schedule

Time
Session
Presenter
8:45-
9:00am
Welcome and Introduction
Bryan Plummer
9:00-
9:30am
Keynote 1: Beyond the Main Track: Findings from the Findings
Lun-Wei Ku
9:30-
10:30am
Keynote 2: State of the Review-nion
Nihar Shah
10:30-
11:00am
Coffee Break
11:00am-
12:00pm
Panel Discussion: Benefits of Findings at computer vision conferences and challenges with implementation
Deqing Sun, Lun-Wei Ku, Nihar Shah, Siyu Tang
Moderator: Boqing Gong
12:00-
1:30pm
Lunch
1:30-
2:15pm
Orals - Session 1
  • BLIP-3: A Family of Open Large Multimodal Models
  • Data Leakage in Visual Datasets
  • DEARLi: Decoupled Enhancement of Recognition and Localization for Semi-supervised Panoptic Segmentation
  • Emergent Visual Grounding in Large Multimodal Models Without Grounding Supervision
  • LOCAL: Latent Orthonormal Contrastive Learning for Paired Image Classification
ICCV'25 Findings Authors
Session Chair: Kate Saenko
2:15-
3:15pm
Poster Session 1 (Exhibit Hall II 95-107) and Coffee Break (from 3pm)
ICCV'25 Findings Authors
3:15-
4:00pm
Orals - Session 2
  • SynBalance: Harnessing Synthetic Data in Long-tailed Recognition
  • DNF-Avatar: Distilling Neural Fields for Real-time Animatable Avatar Relighting
  • Similarity-Aware Selective State-Space Modeling for Semantic Correspondence
  • MIDAS: Modeling Ground-Truth Distributions with Dark Knowledge for Domain Generalized Stereo Matching
ICCV'25 Findings Authors
Session Chair: Kevin Shih
4:00-
5:00pm
Poster Session 2 (Exhibit Hall II 95-107)
ICCV'25 Findings Authors


Organizers

Margrit Betke

Yonatan Bisk

Juan C. Caicedo

Professor, Boston University
Assistant Professor, Carnegie Mellon University
Assistant Professor, University of Wisconsin-Madison

Grigorios Chrysos

Trevor Darrell

Deepti Ghadiyaram

Assistant Professor, University of Wisconsin-Madison
Professor, UC Berkeley
Assistant Professor, Boston University

Boqing Gong

Derek Hoiem

Ziwei Liu

Assistant Professor, Boston University
Professor, University of Illinois at Urbana-Champaign
Associate Professor, Nanyang Technological University

Bryan Plummer

Anna Rohrbach

Bryan Russell

Assistant Professor, Boston University
Professor, TU Darmstadt
Senior Research Scientist, Adobe Research

Kate Saenko

Humphrey Shi

Kevin Shih

Professor, Boston University
Associate Professor, Georgia Tech
Senior Research Scientist, NVIDIA

Contact Info


E-mail: iccv25.findings@gmail.com

Header image credits: Gemini 2.0: "Photorealistic image of machine vision system that is finding new scientific data out in the beaches of Honolulu in Hawaii."

Website based on https://vto-cvpr24.github.io/.