딥 러닝 기반 휴먼 모션 디노이징
Deep Learning-Based Human Motion Denoising
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
본 논문에서는 어텐션 기법을 적용한 양방향 순환신경망을 이용하여 새로운 휴먼 모션 디노이징 방법을 제안한다. 본 방법을 이용하면, 단일 3D 깊이 센서 카메라에서 캡처된 노이즈가 포함된 사람의 움직임이 잘 교정된 자연스러운 움직임으로 자동 조정된다. 양방향 순환신경망에 어텐션 기법을 도입하면, 입력으로 들어온 움직임을 인코딩할 때 여러 자세 중에 더 중요한 자세가 있는 프레임에 더 높은 어텐션 가중치를 부여함으로써, 다른 딥 러닝 네트워크와 비교해 더 나은 최적화 결과와더 높은 정확도를 보인다. 실험을 통해 본 논문에서 제시한 방법이 다양한 스타일의 움직임과 노이즈를 효과적으로 처리함을확인하였으며, 제시한 방법은 모션 캡처 후처리 단계의 애플리케이션으로 충분히 사용 가능할 것으로 기대된다.
Abstract
In this paper, we propose a novel method of denoising human motion using a bidirectional recurrent neural network(BRNN) with an attention mechanism. The corrupted motion captured from a single 3D depth sensor camera isautomatically fixed in the well-established smooth motion manifold. Incorporating an attention mechanism into BRNNachieves better optimization results and higher accuracy than other deep learning frameworks because a higher weightvalue is selectively given to a more important input pose at a specific frame for encoding the input motion. Experimentalresults show that our approach effectively handles various types of motion and noise, and we believe that our methodcan sufficiently be used in motion capture applications as a post-processing step after capturing human motion.
참고문헌 (21)
- Jinxiang Chai, "Performance animation from low-dimensional control signals", ACM Transactions on Graphics, 24(3), pp.686~696 (2005), Association for Computing Machinery (ACM)[DOI]
- Y. Lee, "Motion fields for interactive character locomotion", ACM Trans. Graph, 29(6) (2010)
- Charles F. Rose III, "Artist-Directed Inverse-Kinematics Using Radial Basis Function Interpolation", Computer Graphics Forum, 20(3), pp.239~250 (2001), Wiley[DOI]
- Tomohiko Mukai, "Geostatistical motion interpolation", ACM Transactions on Graphics, 24(3), pp.1062~1070 (2005), Association for Computing Machinery (ACM)[DOI]
- N. D. Lawrence, "Gaussian process latent variable models for visualisation of high dimensional data", Proceedings of the 16th International Conference on Neural Information Processing Systems (NIPS'03), pp.329~336 (2004)
- Keith Grochow, "Style-based inverse kinematics", ACM Transactions on Graphics, 23(3), pp.522~531 (2004), Association for Computing Machinery (ACM)[DOI]
- A. Krizhevsky, "Advances in neural information processing systems", pp.1097~1105 (2012)
- D. Cireşan, "Multi-column deep neural networks for image classification", arXiv preprint arXiv：1202.2745, ~ (2012)
- A. Karpathy, "Large-Scale Video Classification with Convolutional Neural Networks", The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.1725~1732 (2014)
- F. Nasse, "Face detection using gpu-based convolutional neural networks", International Conference on Computer Analysis of Images and Patterns, pp.83~90 (2009)
- Shuiwang Ji, "3D Convolutional Neural Networks for Human Action Recognition", IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(1), pp.221~231 (2013), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Jialue Fan, "Human Tracking Using Convolutional Neural Networks", IEEE Transactions on Neural Networks, 21(10), pp.1610~1623 (2010), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Ossama Abdel-Hamid, "Convolutional Neural Networks for Speech Recognition", IEEE, 2014, pp.22(10), 1533~1545[DOI]
- D. Holden, "Learning motion manifolds with convolutional autoencoders", SIGGRAPH Asia 2015 Technical Briefs (2015)
- A. Pandey, "TCNN: Temporal Convolutional Neural Network for Real-Time Speech Enhancement in The Time Domain", 2019IEEE International Conference on Acoustics, Speech and Signal Processing, pp.6875~6879 (2019)
- C. Lea, "Temporal convolutional networks for action segmentation and detection", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.156~165 (2017)
- L. Sun, "Human Action Recognition Using Factorized Spatio-Temporal Convolutional Networks", 2015IEEE International Conference on Computer Vision (ICCV), pp.4597~4605 (2015)
- J. Guo, "Building an end-to-end spatial-temporal convolutional network for video super-resolution", Thirty-First AAAI Conference on Artificial Intelligence (2017)
- "Carnegie-Mellon Motion Capture Database"[URL]
- S. R. Buss, "Introduction to inverse kinematics with jacobian transpose, pseudoinverse and damped least squares methods", IEEE Journal of Robotics and Automation, 17(16), pp.1~19 (2004)
- S. U. Kim, "Human Motion Denoising Using Attention-Based Bidirectional Recurrent Neural Network", SIGGRAPH Asia 2019 Posters (SA '19) (2019)