AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Scale-Space Hypernetworks for Efficient Biomedical Image Analysis

Jose Javier Gonzalez Ortiz, John Guttag, Adrian V Dalca

NeurIPS 2023 · Architectures and efficiency › CNN design · approach: Compact CNN architectures · 0 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 Fixed U-Net with SSHN hypernetworkU-Net · CAMUS testDice (%): 83.5 -> 86.2Improved
replace Fixed U-Net with SSHN hypernetworkU-Net · OASIS testDice (%): 89.1 -> 90.7Improved
replace Fixed U-Net with SSHN hypernetworkU-Net · PanDental testDice (%): 88.9 -> 89.9Improved
replace Fixed U-Net with SSHN hypernetworkU-Net · WBC testDice (%): 90.9 -> 92.7Improved
use stochastic rescaling factor during training of a plain U-Net instead of fixedU-Net · OASIS testDice (%): 89.1 -> 84.8Hurt
add FiLM conditioning on stochastically sampled rescaling factor instead of fixedU-Net · OASIS testDice (%): 89.1 -> 87.2Hurt
stochastic rescaling vs fixed on CAMUSU-Net · CAMUS testDice (%): 83.5 -> 83.3No clear effect
widen the training prior of phi from r=0 (constant 0.5) to wider uniform range (ablation)SSHN · OASIS test, evaluated at phi=0.5DiceImproved
train SSHN with r = 0 (constant phi = 0.5) (ablation)SSHN · OASIS test, phi=0.5DiceHurt
widen prior to U(0.4, 0.6) beyond narrow ranges (ablation)SSHN · OASIS testDiceNo clear effect
reduce rescaling factor from 0.5 to 0.4SSHN · OASIS and PanDentalinference FLOPs: over 50% reductionImproved
reduce rescaling factor from 0.5 to 0.3SSHN · OASIS and PanDentalinference FLOPs: 70% reductionImproved

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