AI Paper Atlas

Research map / Video / Video segmentation and tracking

Fast Window-Based Event Denoising With Spatiotemporal Correlation Enhancement

Huachen Fang, Jinjian Wu, Qibin Hou, Weisheng Dong, Guangming Shi

TPAMI 2024 · Video › Video segmentation and tracking · approach: Event-based video reconstruction · 11 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 BEC unit (bone events check) instead of farthest sampling (ablation)WedNet · DVSCLEANSNR: 21.65 -> 25.73Improved
add BEC unit instead of farthest sampling (ablation)WedNet · DVSNOISE20RPMD: 35.49 -> 28.12Improved
add BEC unit instead of farthest sampling (ablation)WedNet · ED-KoGTLSNR: 22.31 -> 26.97Improved
increase iterative number in SSFE module (ablation)WedNet · DVSCLEAN, DVSNOISE20, ED-KoGTLSNR / RPMDNo clear effect
increase sampling event number in first sampling operation beyond sufficient level (ablation)WedNet · DVSCLEAN, DVSNOISE20, ED-KoGTLSNR / RPMDMixed
use SSFE module instead of Spatial Feature Embedding or Res-block (ablation)WedNet · DVSCLEAN, DVSNOISE20, ED-KoGTLSNR / RPMDImproved
use WedNet instead of AEDNetAEDNet · Denoising runtime on DVSNOISE20time (s): 8379.58 -> 427.54Improved
use WedNet instead of AEDNet on 100% noise ratioAEDNet · DVSCLEAN (100% noise)SNR: 25.08 -> 24.65Hurt
use WedNet instead of AEDNetAEDNet · ED-KoGTL (average)SNR: 25.54 -> 26.97Improved
denoise events with WedNet before HATS classificationHATS · MNIST-DVSaccuracy: 98.4% -> 99.1%Improved
denoise events with WedNet before HATS classificationHATS · N-CARSaccuracy: 90.2% -> 92.1%Improved
denoise events with WedNet before HATS classificationHATS · CIFAR10-DVSaccuracy: 52.4% -> 60.1%Improved

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