AI Paper Atlas

Research map / 3D vision / Depth and stereo

Deep Depth from Focal Stack with Defocus Model for Camera-Setting Invariance

Yuki Fujimura, Masaaki Iiyama, Takuya Funatomi, Yasuhiro Mukaigawa

IJCV 2023 · 3D vision › Depth and stereo · approach: Depth from light field and defocus · 23 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 cost volume from defocus model instead of direct focal stack inputOurs vs DefocusNet · NYU Depth V2 blurred (train DefocusNet)RMSE: 0.621 -> 0.186Improved
use cost volume under different train/test focus distancesOurs vs DefocusNet · DefocusNet datasetRMSE: 0.299 -> 0.242Improved
add normalization to cost volume (ablation)Ours (outlier removal only) · DefocusNet dataset sigma=1.0RMSE: 0.261 -> 0.239Improved
add outlier removal to normalized cost volume (ablation)Ours (normalization only) · DefocusNet dataset sigma=9.0RMSE: 0.422 -> 0.403Improved
add outlier removal to normalized cost volume (ablation)Ours (normalization only) · DefocusNet dataset sigma=1.0RMSE: 0.232 -> 0.239Hurt
add normalization to cost volume at large depth scale (ablation)Ours (outlier removal only) · DefocusNet dataset sigma=9.0RMSE: 0.475 -> 0.403Improved
compensate scale by rescaling baseline outputsDefocusNet · NYU Depth V2 blurredRMSE: 0.621 -> 0.322Improved
train on NYU Depth V2 instead of DefocusNet datasetOurs · NYU Depth V2 blurredRMSE: 0.186 -> 0.054Improved
use explicit defocus model on DefocusNet dataset where defocus cues are short-rangeOurs vs AiFDepthNet · DefocusNet datasetRMSE: 0.156 -> 0.239Hurt
train and test on NYU Depth V2 with matched camera settingsOurs vs DefocusNet · NYU Depth V2 blurred (train NYU)RMSE: 0.029 -> 0.054Hurt
add neighboring pixels to standard deviation costOurs · real captured datadepth noiseImproved

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