atlas

Research map / Architectures and efficiency / CNN design

LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation

CVPR 2018 · Architectures and efficiency › CNN design · approach: Efficient CNN scaling and dynamic inference · 799 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 feature warping (ablation)LiteFlowNet-pre · FlyingChairs, Sintel clean/final, KITTI12/15 trainAEE: 3.75 -> 2.98; 2.70 -> 1.82; 4.70 -> 3.54; 4.17 -> 2.90; 9.22 -> 4.83; 18.24 -> 12.32Helped
add descriptor matching (ablation)LiteFlowNet-pre · KITTI12/15 trainAEE: 5.03 -> 4.83; 13.20 -> 12.32Helped
add sub-pixel refinement (ablation)LiteFlowNet-pre · FlyingChairs, Sintel clean/final, KITTI12/15 trainAEE: 2.98 -> 1.82; 3.54 -> 2.90; 4.81 -> 4.45; 6.17 -> 4.83; 14.52 -> 12.32Helped
add feature-driven local-convolution regularization (ablation)LiteFlowNet-pre · FlyingChairs, Sintel clean/final, KITTI12/15 trainAEE: 1.82 -> 1.57; 2.90 -> 2.78; 4.45 -> 4.17; 4.83 -> 4.56; 12.32 -> 11.58Helped
train on Chairs followed by Things3D rather than Chairs aloneLiteFlowNet · Sintel clean/final and KITTI12/15 trainAEE: 2.78 -> 2.48; 4.17 -> 4.04; 4.56 -> 4.00; 11.58 -> 10.39Helped

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