Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Visual Sync: Multi‑Camera Synchronization via Cross‑View Object Motion
Authors: Shaowei Liu, David Yao, Saurabh Gupta, Shenlong Wang
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on four diverse, challenging datasets show that Visual Sync outperforms baseline methods, achieving an average synchronization error below 130 ms. We validate our approach on diverse datasets and show strong performance across different scenes, motions, and camera setups, and achieve high-precision synchronization even under severe viewpoints. Specifically, we outperform Sync Nerf [26], a recent method for this task by radiance field optimization, and adaptations of two recent methods Uni4D [62] and MAST3R [30]. |
| Researcher Affiliation | Academia | Shaowei Liu1 David Yifan Yao1 Saurabh Gupta1 Shenlong Wang1 1University of Illinois Urbana-Champaign |
| Pseudocode | No | The paper describes a three-stage optimization strategy (Stage 0: Visual Cue Extraction, Stage 1: Estimating Pairwise Offsets, Stage 2: Global Offset Esimation) but does not present these stages in a structured pseudocode or algorithm block. |
| Open Source Code | No | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [No] Justification: All data we used are public. We mainly use off-the-shelf models in our paper, we will release the complete code with instructions upon acceptance. |
| Open Datasets | Yes | Datasets. We evaluate our method on a comprehensive suite of multi-view video datasets capturing dynamic scenes. [...] CMU Panoptic [22] features a real-world indoor dataset... Egohumans [25] is a challenging multi-view egocentric and static cameras... 3DPOP [37] is a large scale 2D to 3D posture, identity and trajectory dataset... Unsynchronized Dynamic Blender Dataset (UDBD) is a synthetic toy example created with dynamic blender assets used in Sync Ne RF [26]. |
| Dataset Splits | No | To prepare these multi-view datasets for multi-video synchronization, we take subsequences of each video while ensuring that they all have a common overlap. Each sequence is roughly 10 seconds long, with a random cropping of around 2-3 seconds from the front and back to simulate offsets and unsynchronized videos. These offsets are used for evaluative purposes. |
| Hardware Specification | Yes | All runtimes are measured on a single A6000 GPU. |
| Software Dependencies | No | The paper mentions several tools used such as VGGT [56], MAST3R [30], Co Tracker3 [23], GPT4o, SAM, DEVA [9], and Huber loss for optimization, but it does not specify version numbers for any of these software components or libraries. |
| Experiment Setup | Yes | Implementation details. Given the input videos, we first extract dynamic object categories using GPT [1] and apply Grounded-SAM [43, 30] to obtain initial per-frame segmentations. We then run DEVA [9] to track these instance masks across time, producing temporally consistent segmentations for each moving object. For each tracked instance, we apply Co Tracker3 [23] to perform per-instance temporal tracking. To establish cross-view correspondences, we sample keyframes every 10 frames and query Mast3R [30] within the dynamic instance masks, linking tracklets across views. Camera poses are estimated by VGGT [56]. We compute pairwise synchronization energy over the maximum overlapping time window between video pairs and evaluate energy across different offsets. Finally, we select reliable pairs for global synchronization. |