AI Paper Atlas

Research map / 3D vision / Depth and stereo

Content-Aware Unsupervised Deep Homography Estimation

Jirong Zhang, Chuan Wang, Shuaicheng Liu, Lanpeng Jia, Nianjin Ye, Jue Wang, Ji Zhou, Jian Sun

ECCV 2020 · 3D vision › Depth and stereo · approach: Self-supervised monocular depth

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 content-aware mask (attention and loss weighting) (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.38 -> 1.82Improved
use mask as attention map and loss weight rather than attention only (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.27 -> 1.82Improved
use mask as attention map in addition to loss weighting (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.07 -> 1.82Improved
add triplet loss term (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.84 -> 1.82Improved
add feature extractor instead of raw pixel intensity (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.03 -> 1.82Improved
use ResNet-34 instead of VGG backbone (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.17 -> 1.82Improved
use ResNet-34 instead of ResNet-18 backbone (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.06 -> 1.82No clear effect
use ResNet-34 instead of ShuffleNet-v2 backbone (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.44 -> 1.82Improved
add two-stage training (attention disabled first, then enabled) (ablation)Content-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 1.90 -> 1.82Improved
use proposed method instead of SIFT + RANSACContent-aware homography network · Homography dataset, LT low texture scenesmean l2 error (pixels): 2.56 -> 1.90Improved
use proposed method instead of SIFT + MAGSACContent-aware homography network · Homography dataset, RE regular scenesmean l2 error (pixels): 1.71 -> 1.81Hurt
use proposed method instead of LIFT + MAGSACContent-aware homography network · Homography dataset, all scenes avgmean l2 error (pixels): 2.07 -> 1.82Improved

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