AI Paper Atlas

Research map / Video / Video segmentation and tracking

C2ST: Cross-Modal Contextualized Sequence Transduction for Continuous Sign Language Recognition

Huaiwen Zhang, Zihang Guo, Yang Yang, Xin Liu, De Hu

ICCV 2023 · Video › Video segmentation and tracking · approach: Sign language and physiological video analysis · 29 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 cross-modal context learning (gloss language model fused with visual features) (ablation)C2ST (TLP + Swin-T baseline) · Phoenix-2014WER test (%): 19.9 -> 19.0Improved
add contextualized sequence transduction loss (ablation)C2ST (TLP + Swin-T baseline) · Phoenix-2014WER test (%): 19.9 -> 19.1Improved
use both CCL and CST together (ablation)C2ST · Phoenix-2014WER test (%): 19.9 -> 17.7Improved
replace CTC loss with conditional gloss decoder loss (CGD) (ablation)C2ST · Phoenix-2014WER test (%): 19.0 -> 18.3Improved
add sequence-level transduction calibration to CGD loss (ablation)C2ST · Phoenix-2014WER test (%): 18.3 -> 17.7Improved
add video-level fusion on top of chunk-level fusion (ablation)C2ST · Phoenix-2014WER test (%): 18.8 -> 17.7Improved
chunk-level fusion alone (ablation)C2ST · Phoenix-2014WER test (%): 19.1 -> 18.8No clear effect
use BERT-Base instead of GloVe or word embedding as gloss language model (ablation)C2ST · Phoenix-2014WER test (%): 18.2 (GloVe) -> 17.7Improved
use larger Swin-S backbone instead of Swin-T (ablation)C2ST · Phoenix-2014WER test (%): 17.7 -> 18.2Hurt
use Swin-T instead of ResNet18 backbone (ablation)C2ST · Phoenix-2014WER test (%): 18.3 -> 17.7Improved
adopt Swin and BERT in C2ST instead of ResNet and word embeddingC2ST · Phoenix-2014 efficiencyParams / throughput: 60.6M -> 187.8M params; 15.3 -> 3.0 infer throughputHurt

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