AI Paper Atlas

Research map / Detection and segmentation / CNN features

Efficient Segmentation: Learning Downsampling Near Semantic Boundaries

Dmitrii Marin, Zijian He, Peter Vajda, Priyam Chatterjee, Sam Tsai, Fei Yang, Yuri Boykov

ICCV 2019 · Detection and segmentation › CNN features · approach: Context modeling for semantic segmentation · 112 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
replace uniform downsampling with learned boundary-driven non-uniform downsamplingsegmentation network · ApolloScape val (32x32, target classes)mIoU: 0.19 -> 0.24Improved
replace uniform downsampling with adaptive downsamplingsegmentation network · ApolloScape val (512x512, target classes)mIoU: 0.44 -> 0.47Improved
add adaptive downsamplingPSP-Net · CityScapes val (64x64)mIoU: 0.29 -> 0.32Improved
add adaptive downsampling at higher resolutionPSP-Net · CityScapes val (256x256)mIoU: 0.54 -> 0.54No clear effect
add adaptive downsamplingDeeplabv3+ · CityScapes val (160x160)mIoU: 0.54 -> 0.58Improved
add adaptive downsamplingsegmentation network · Synthia (target classes)mIoU: improvement of 1.5% to 3% at negligible costImproved
add adaptive downsamplingsegmentation network · Synthia (32x32, all classes)mIoU: 0.65 -> 0.67Improved
add adaptive downsamplingsegmentation network · Supervisely person (16x16)mIoU: 0.69 -> 0.73Improved
add adaptive downsamplingsegmentation network · Supervisely person (32x32)mIoU: 0.76 -> 0.76No clear effect
focus sampling on boundaries of selected target classessegmentation network · ApolloScape (non-target classes)IoU per class: non-target classes sometimes lower, e.g. sidewalk 0.29 -> 0.38 but pole 0.11 -> 0.08 at 32Mixed
adaptive downsampling on small objectssegmentation network · ApolloScape recall by object size (64x64)object recallImproved
lower input resolution to 64x64 with adaptive downsamplingsegmentation network · Supervisely trimap accuracyaccuracy near boundariesImproved

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