수화 인식을 위한 다양한 액션 모델의 평가와 분석
Translating the Silent Voice: A Benchmark of Existing Action Models for Sign Language Recognition
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Various different action recognition models have been used for Sign Language Recognition (SLR). Our research questions are 1) why a particular model was chosen, and 2) what the computational requirements, accuracy, and trade-offs of each architecture are. This paper experimentally studies the performance of different existing action recognition models against two SLR datasets: WLASL2000 and AUTSL. Our findings include one model named Channel-Separated Network (CSN) that achieves the highest accuracy in all benchmarks conducted with results comparable to the current state-of-the-art, SignBERT for WLASL100.
Abstract
Various different action recognition models have been used for Sign Language Recognition (SLR). Our research questions are 1) why a particular model was chosen, and 2) what the computational requirements, accuracy, and trade-offs of each architecture are. This paper experimentally studies the performance of different existing action recognition models against two SLR datasets: WLASL2000 and AUTSL. Our findings include one model named Channel-Separated Network (CSN) that achieves the highest accuracy in all benchmarks conducted with results comparable to the current state-of-the-art, SignBERT for WLASL100.
참고문헌 (26)
- S. Yan, "Spatial temporal graph convolutional networks for skeleton-based action recognition", AAAI Conference on Artificial Intelligence, 32(1) (2018)
- L. Zhang, "Learning spatiotemporal features using 3dcnn and convolutional lstm for gesture recognition", 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), pp.3120~3128 (2017)
- D. Tran, "A closer look at spatiotemporal convolutions for action recognition", 2018 IEEE, 2018, 6450~6459
- C. Neidle, "Resources for Computer-Based Sign Recognition from Video, and the Criticality of Consistency of Gloss Labeling across Multiple Large ASL Video Corpora", LREC2022 10th Workshop on the Representation and Processing of Sign Languages:Multilingual Sign Language Resources, pp.165~172 (2022)
- L. R. Cerna, "A multimodal LIBRAS-UFOP Brazilian sign language dataset of minimal pairs using a microsoft Kinect sensor", Expert Systems with Applications, 167 (2021)[DOI]
- L. -G. Zhang, "Recognition of sign language subwords based on boosted hidden Markov models", 7th international conference on Multimodal interfaces, pp.282~287 (2005)
- O. M. Sincan, "Autsl : A large scale multi-modal turkish sign language dataset and baseline methods", IEEE Access, 8, pp.181340~181355 (2020)[DOI]
- E. Liu, "Weighted Multi-modal Sign Language Recognition", 2023 Australasian Conference on Robotics and Automation (ACRA 2023) (2023)
- M. Boháček, "Sign pose-based transformer for wordlevel sign language recognition", 2022 IEEE, 2022, 182~191
- M. De Coster, "Isolated sign recognition from rgb video using pose flow and selfattention", 2021 IEEE, 2021, 3441~3450
- "OpenMMLab’s Next Generation Video Understanding Toolbox and Benchmark"[URL]
- Y. Jing, "Learning Spatiotemporal Features of CSI for Indoor Localization With Dual-Stream 3D Convolutional Neural Networks", IEEE Access, 7, pp.147571~147585 (2019)[DOI]
- L. Wang, "Temporal segment networks : Towards good practices for deep action recognition", Computer Vision – ECCV 2016, pp.20~36 (2016)
- Z. Liu, "Tam : Temporal adaptive module for video recognition", 2021 IEEE, 2021, 13708~13718
- D. Tran, "Video classification with channel-separated convolutional networks", 2019IEEE, 2019, 5552~5561
- H. Shao, "Temporal interlacing network", AAAI Conference on Artificial Intelligence, 34(07), pp.11966~11973 (2020)