AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

SubUNets: End-To-End Hand Shape and Continuous Sign Language Recognition

Necati Cihan Camgoz, Simon Hadfield, Oscar Koller, Richard Bowden

ICCV 2017 · Architectures and efficiency › RNNs and normalization · approach: Differentiable optimization layers · 382 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 a second BLSTM layer (ablation)Word SubUNet · PHOENIX-2014 testWER (%): 43.8 -> 43.1Improved
add full frames input instead of hand patches only (ablation)Word SubUNet · PHOENIX-2014 testWER (%): 44.5 -> 43.1Improved
train combined SubUNets end-to-end instead of fixing pretrained weights (ablation)Hand + Full Frame SubUNet · PHOENIX-2014 testWER (%): 44.2 -> 42.1Improved
add Hand SubUNet hand shape CTC to Full Frame Word SubUNet (ablation)Word SubUNet · PHOENIX-2014 testWER (%): 43.1 -> 42.1Improved
use hand shape expert CTC instead of generic word CTC (ablation)combined SubUNet · PHOENIX-2014 testWER (%): 42.6 -> 42.1Improved
use hand shape expert CTC instead of generic word CTC (ablation)combined SubUNet · PHOENIX-2014 devWER (%): 43.0 -> 43.1No clear effect
use Full Sum decoding instead of Viterbi (ablation)combined SubUNet · PHOENIX-2014 testWER (%): 42.7 -> 42.1Improved
add HMM topology with language model decoding instead of CTC topology (ablation)combined SubUNet · PHOENIX-2014 devWER (%): 43.1 -> 40.8Improved
add LM topology with silence states instead of CTC topology (ablation)combined SubUNet · PHOENIX-2014 testWER (%): 42.1 -> 42.2No clear effect
add Hand SubUNet end-to-end alignment instead of iterative realignmenthand shape recognition · One-Million HandsTop-1 accuracy (%): 62.8 -> 80.3Improved
use CaffeNet instead of ResNet-50 as spatial encoder (ablation)dummy SubUNet · GPU memory capacitymax frames loadable: 35 -> 1450Improved

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