AI Paper Atlas

Research map / 3D vision / Depth and stereo

Depth Estimation via Affinity Learned with Convolutional Spatial Propagation Network

Xinjing Cheng, Peng Wang, Ruigang Yang

ECCV 2018 · 3D vision › Depth and stereo · approach: Dense depth prediction and completion · 372 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 mirror connections (UNet) to the depth network, concatenating after reluMa et al. network · NYU v2 (color only)RMSE: 0.230 -> 0.137Improved
replace SPN with CSPN for refinementMa et al. network · NYU v2 (color only)RMSE: 0.172 -> 0.162Improved
replace SPN with CSPN with sparse depth embeddedUNet · NYU v2 (500 sparse samples, preserve SD)RMSE: 0.144 -> 0.117Improved
add bilateral filtering post-processingMa et al. network · NYU v2 (color only)RMSE: 0.230 -> 0.479Hurt
add UNet mirror-connection network to SPN with sparse depthSPN with sparse depth · NYU v2 (500 sparse samples, preserve SD)RMSE: 0.144 vs 0.137No clear effect
add CSPN with sparse depth to the UNet networkUNet · KITTI (500 sparse samples)delta<1.02 (%): 62.6 -> 70.2Improved
add SPN with sparse depth to the UNet networkUNet · KITTI (500 sparse samples)RMSE: 3.049 -> 3.248Hurt
concatenate mirror connection after relu instead of after conv (ablation)UNet mirror connection · NYU v2 (color only)RMSE: 0.531 -> 0.137Improved
concatenate mirror connection after relu instead of after bn (ablation)UNet mirror connection · NYU v2 (color only)RMSE: 0.158 -> 0.137Improved
increase CSPN iterations up to 24 (kernel 3) (ablation)CSPN · NYU v2RMSE: accuracy is saturated when the number of iterations is increased to 24Improved
increase CSPN kernel size beyond 7 to 9 (N=12) (ablation)CSPN · NYU v2RMSE: saturated at size of 7; performance drops slightly at 9Hurt
use CSPN instead of SPN for propagation speedrefinement module · 1024x768 image, 4 iterationsms: 127.902 -> 3.689Improved

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