Meaningful Pose-Based Sign Language Evaluation
Not on the map. Why?
Abstract (arXiv)
We present a comprehensive study on meaningfully evaluating sign language utterances in the form of human skeletal poses. The study covers keypoint distance-based, embedding-based, and back-translation-based metrics. We show tradeoffs between different metrics in different scenarios through automatic meta-evaluation of sign-level retrieval and a human correlation study of text-to-pose translation across different sign languages. Our findings and the open-source pose-evaluation toolkit provide a practical and reproducible way of developing and evaluating sign language translation or generation systems.
Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Builds on 38 of its 72 references that are in the map. First posted on arXiv 8 Oct 2025.
Builds on
38 of 72 references.
References · 38
- 2019BERTScore: Evaluating Text Generation with BERTcited by 5,512
- 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,424
- 2018A Call for Clarity in Reporting BLEU Scorescited by 1,799
- 2019Expressive Body Capture: 3D Hands, Face, and Body from a Single Imagecited by 1,612
- 2020BLEURT: Learning Robust Metrics for Text Generationcited by 1,203
- 2018OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fieldscited by 1,162
- 2022MotionDiffuse: Text-Driven Human Motion Generation with Diffusion Modelcited by 780
- 2019MediaPipe: A Framework for Building Perception Pipelinescited by 583
- 2022MotionCLIP: Exposing Human Motion Generation to CLIP Spacecited by 408
- 2020Deep Learning-Based Human Pose Estimation: A Surveycited by 207
- 2021Improving Sign Language Translation with Monolingual Data by Sign Back-Translationcited by 179
- 2020How2Sign: A Large-scale Multimodal Dataset for Continuous American Sign Languagecited by 173
26 more
- 2020Progressive Transformers for End-to-End Sign Language Productioncited by 120
- 2023Gloss-free Sign Language Translation: Improving from Visual-Language Pretrainingcited by 105
- 2019Sign Language Recognition, Generation, and Translation: An Interdisciplinary Perspectivecited by 87
- 2023YouTube-ASL: A Large-Scale, Open-Domain American Sign Language-English Parallel Corpuscited by 75
- 2021Including Signed Languages in Natural Language Processingcited by 68
- 2020Adversarial Training for Multi-Channel Sign Language Productioncited by 57
- 2021Continuous 3D Multi-Channel Sign Language Production via Progressive Transformers and Mixture Density Networkscited by 50
- 2021Mixed SIGNals: Sign Language Production via a Mixture of Motion Primitivescited by 50
- 2023Neural Sign Actors: A diffusion model for 3D sign language production from textcited by 49
- 2024T2S-GPT: Dynamic Vector Quantization for Autoregressive Sign Language Production from Textcited by 30
- 2023SignAvatars: A Large-scale 3D Sign Language Holistic Motion Dataset and Benchmarkcited by 29
- 2021Data Augmentation for Sign Language Gloss Translationcited by 28
- 2022Considerations for meaningful sign language machine translation based on glossescited by 25
- 2022Ham2Pose: Animating Sign Language Notation into Pose Sequencescited by 21
- 2023What is the Best Automated Metric for Text to Motion Generation?cited by 19
- 2024A Simple Baseline for Spoken Language to Sign Language Translation with 3D Avatarscited by 19
- 2022Machine Translation between Spoken Languages and Signed Languages Represented in SignWritingcited by 18
- 2023DUB: Discrete Unit Back-translation for Speech Translationcited by 18
- 2023The Sem-Lex Benchmark: Modeling ASL Signs and Their Phonemescited by 17
- 2023Improving 3D Pose Estimation for Sign Languagecited by 16
- 2022COMET-QE and Active Learning for Low-Resource Machine Translationcited by 10
- 2023An Open-Source Gloss-Based Baseline for Spoken to Signed Language Translationcited by 9
- 2023Autoregressive Sign Language Production: A Gloss-Free Approach with Discrete Representationscited by 9
- 2022Sockeye 3: Fast Neural Machine Translation with PyTorchcited by 3
- 2023pose-format: Library for Viewing, Augmenting, and Handling .pose Filescited by 3
- 2024Modelling the Distribution of Human Motion for Sign Language Assessmentcited by 3
Built on by
Citing papers per year: 2026: 6.
6 papers.
- 2026M3T: Discrete Multi-Modal Motion Tokens for Sign Language Productioncited by 2
- 2026BackTranslation2.0 -- A Linguistically Motivated Metric to Assess Sign Language Productioncited by 1
- 2026Evaluation of Pose Estimation Systems for Sign Language Translationcited by 0
- 2026Direct Translation between Sign Languagescited by 0
- 2026Conditional Collapse in Sign Language Production: A Diagnostic and a Scaling Argumentcited by 0
- 2026Beyond BLEU: A Case for Redefining Sign Language Translation Benchmarkscited by 0
- 2026Beyond BLEU: A Case for Redefining Sign Language Translation Benchmarks3 Sep 2026
- 2026BackTranslation2.0 -- A Linguistically Motivated Metric to Assess Sign Language Production27 Jun 2026
- 2026Conditional Collapse in Sign Language Production: A Diagnostic and a Scaling Argument1 Jun 2026
- 2026Direct Translation between Sign Languages20 May 2026
- 2026Evaluation of Pose Estimation Systems for Sign Language Translation27 Apr 2026
- 2026M3T: Discrete Multi-Modal Motion Tokens for Sign Language Production24 Mar 2026