AI Paper Atlas

Research map / Video / Action recognition

End-to-End United Video Dehazing and Detection

Boyi Li, Xiulian Peng, Zhangyang Wang, Jizheng Xu, Dan Feng

AAAI 2018 · Video › Action recognition · approach: Fully supervised temporal action detection · 130 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
Use five-frame K-level fusion at conv2 rather than the strongest listed image-based dehazing baselineAOD-Net on TestSet V1 · TestSet V1PSNR: 20.6828 -> 20.9908Improved
Use five-frame K-level fusion at conv2 rather than the strongest listed baseline for SSIMCAP on TestSet V1 · TestSet V1SSIM: 0.8848 -> 0.9087Improved
Jointly train video dehazing and video detection instead of using the best listed detection baselineJAOD-Faster R-CNN on TestSet V2 · TestSet V2MAP: 0.758 -> 0.802Improved
Fuse five-frame inputs at conv2 rather than at the input or output level (ablation)EVD-Net · TestSet V1PSNR: I-level fusion, 5 frames: 20.6095; J-level fusion, 5 frames: 20.3675; K-level fusion, conv2, 5 frames: 20.9908Improved
Use five rather than three frames for K-level fusion at conv2 (ablation)EVD-Net · TestSet V1PSNR: 20.6998 -> 20.9908Improved
Use seven or nine rather than five frames for K-level fusion at conv2 (ablation)EVD-Net · TestSet V1PSNR: 20.9908 -> 20.7901; 20.9908 -> 20.7355Hurt
Move the K-level fusion point from conv2 to conv3 while using five frames (ablation)EVD-Net · TestSet V1PSNR: 20.9908 -> 20.7780Hurt
Move the K-level fusion point from conv2 to conv1 while using five frames (ablation)EVD-Net · TestSet V1PSNR: 20.9908 -> 20.8240Hurt
Move the K-level fusion point from conv2 to conv4 or conv5(K) while using five frames (ablation)EVD-Net · TestSet V1PSNR: 20.9908 -> 20.6756; 20.9908 -> 20.7942Hurt

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