AI Paper Atlas

Research map / Detection and segmentation / CNN features

Submodular Field Grammars: Representation, Inference, and Application to Image Parsing

Abram L. Friesen, Pedro M Domingos

NeurIPS 2018 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 4 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
parse with a generated submodular field grammar instead of using DeepLab features aloneDeepLab · SBD test, 143 imagesmean pixel accuracy: 87.77 -> 90.03Improved
replace a flat submodular MRF with an SFG over DeepLab featuresDeepLab · SBD test, 143 imagesmean pixel accuracy: 87.93 -> 90.03Improved
add a flat submodular MRF on top of DeepLab featuresDeepLab · SBD test, 143 imagesmean pixel accuracy: 87.77 -> 87.93No clear effect
use SFG-Parse instead of alpha-expansion and BP for MAP inference as boundary strength variesSFG · synthetic SFGs, boundary scale factorminimum energy and inference timeImproved
increase grammar height from SFG-Parse versus alpha-expansion and BPSFG-Parse · synthetic SFGs, wBF 20inference timeImproved
increase number of productions per nonterminal, SFG-Parse versus alpha-expansion and BPSFG-Parse · synthetic SFGs, wBF 20energy and inference timeImproved

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