AI Paper Atlas

Research map / Detection and segmentation / CNN features

The Devil is in the Decoder: Classification, Regression and GANs

Zbigniew Wojna, Vittorio Ferrari, Sergio Guadarrama, Nathan Silberman, Liang-Chieh Chen, Alireza Fathi, Jasper Uijlings

IJCV 2019 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 86 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 residual-like connections to decoders (ablation)decoders across tasks · seven pixel-wise taskstask metricsImproved
add residual-like connections (ablation)stacked hourglass decoder · Human keypoints estimationPCKhNo clear effect
add residual connections to transposed convolution decoder (ablation)Transposed · Depth regression artifacts, x-axisartifact rate: 27.6% -> 23.1%Improved
add skip connections to transposed convolution decoder (ablation)Transposed · Depth regression artifacts, x-axisartifact rate: 27.6% -> 30.6%Hurt
add skip connections to decomposed transposed decoder (ablation)Dec. Transposed · Depth regression artifacts, x-axisartifact rate: 27.0% -> 33.4%Hurt
add skip connections to bilinear upsampling + conv (ablation)Bilinear Ups. + Conv. · Depth regression artifacts, x-axisartifact rate: 14.3% -> 19.5%Hurt
use bilinear upsampling + conv instead of transposed convolution (ablation)decoder for depth regression · Depth regression artifacts, x-axisartifact rate: 27.6% -> 14.3%Improved
use bicubic instead of nearest neighbour upsampling (ablation)Bilinear Ups. + Conv, res+skip · Human keypoints estimationPCK/0.50: 0.846 -> 0.855No clear effect
use bilinear instead of nearest neighbour upsampling (ablation)Bilinear Ups. + Conv, res+skip · Human keypoints estimationPCK/0.50: 0.846 -> 0.854No clear effect
use bilinear instead of nearest neighbour upsampling (ablation)Bilinear Ups. + Conv, res · GANs on Cifar 10FID: 28.0 -> 29.8Hurt
use bicubic instead of nearest neighbour upsampling (ablation)Bilinear Ups. + Conv, res · GANs on Cifar 10FID: 28.0 -> 28.1No clear effect
use depth-to-space upsampling without residual connections (ablation)decoder · Semantic segmentationmIoUImproved

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