AI Paper Atlas

Research map / Video / Video segmentation and tracking

Efficient Event Camera Data Pretraining with Adaptive Prompt Fusion

Quanmin Liang, Qiang Li, Shuai Liu, Xinzi Cao, Jinyi Lu, Feidiao Yang, Wei Zhang, Kai Huang, Yonghong Tian

ICCV 2025 · Video › Video segmentation and tracking · approach: Event-based video reconstruction · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add overlapping patch embedding (overlap window) (ablation)STP · N-ImageNetfinetuned top-1 accuracy: 67.04 -> 68.87Improved
add fine-grained conv after patch embedding (ablation)STP · N-ImageNetfinetuned top-1 accuracy: 67.36 -> 68.87Improved
add both overlap window and conv (ablation)STP · N-ImageNetpretraining top-1 accuracy: 59.30 -> 66.01Improved
add local temporal aggregation (LTA) (ablation)STP · N-ImageNetfinetuned top-1 accuracy: 68.12 -> 68.87Improved
use STP instead of E2VID reconstruction prompting (ablation)image-pretrained ViT-S · N-ImageNetfinetuned top-1 accuracy: 64.93 -> 68.87Improved
use STP instead of vanilla patch-embedding prompt (ablation)image-pretrained ViT-S · N-ImageNetfinetuned top-1 accuracy: 66.89 -> 68.87Improved
use TECM instead of ECM input (ablation)STP · N-ImageNettop-1 accuracy: 67.82 -> 68.87Improved
use TECM instead of Voxel Grid input (ablation)STP · N-ImageNettop-1 accuracy: 58.26 -> 68.87Improved
use TECM instead of EST input (ablation)STP · N-ImageNettop-1 accuracy: 68.07 -> 68.87Improved
scale frozen backbone from ViT-S/16 to ViT-L/16 (ablation)STP · N-ImageNetfinetuned top-1 accuracy: 68.87 -> 78.06Improved
swap ViT-S/16 backbone for ConvNeXt-T (ablation)STP · N-ImageNetfinetuned top-1 accuracy: 68.87 -> 71.90Improved
swap ViT-S/16 backbone for ResNet50 (ablation)STP · N-ImageNetfinetuned top-1 accuracy: 68.87 -> 65.93Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

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