MobileViT와 전이학습을 활용한 사람 자세 추정 알고리즘의 경량화
Lightening of Human Pose Estimation Algorithm Using MobileViT and Transfer Learning
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
본 논문에서는 매개변수가 더 적고, 빠르게 추정 가능한 MobileViT 기반 모델을 통해 사람 자세 추정 과업을 수행할 수 있는 모델을 제안한다. 기반 모델은 합성곱 신경망의 특징과 Vision Transformer의 특징이 결합한 구조를 통해 경량화된 성능을 입증한다. 본 연구에서 주요 매커니즘이 되는 Transformer는 그 기반의 모델들이 컴퓨터 비전 분야에서도 합성곱 신경망 기반의 모델들대비 더 나은 성능을 보이며, 영향력이 커지게 되었다. 이는 사람 자세 추정 과업에서도 동일한상황이며, Vision Transformer기반의 ViTPose가 COCO, OCHuman, MPII 등 사람 자세 추정 벤치마크에서 모두 최고 성능을 지키고 있는 것이 그 적절한 예시이다. 하지만 Vision Transformer는 매개변수의 수가 많고 상대적으로 많은 연산량을 요구하는 무거운 모델 구조를 가지고 있기 때문에, 학습에 있어 사용자에게 많은 비용을 야기시킨다. 이에 기반 모델은 Vision Transformer가 많은계산량을 요구하는 부족한 Inductive Bias 계산 문제를 합성곱 신경망 구조를 통한 Local Representation으로 극복하였다. 최종적으로, 제안 모델은 MS COCO 사람 자세 추정 벤치마크에서제공하는 Validation Set으로 ViTPose 대비 각각 5분의 1과 9분의 1만큼의 3.28GFLOPs, 972만 매개변수를 나타내었고, 69.4 Mean Average Precision을 달성하여 상대적으로 우수한 성능을 보였다.
Abstract
In this paper, we propose a model that can perform human pose estimation through a MobileViT-based model with fewer parameters and faster estimation. The based model demonstrates lightweight performance through a structure that combines features of convolutional neural networks with features of Vision Transformer. Transformer, which is a major mechanism in this study, has become more influential as its based models perform better than convolutional neural network-based models in the field of computer vision. Similarly, in the field of human pose estimation, Vision Transformer-based ViTPose maintains the best performance in all human pose estimation benchmarks such as COCO, OCHuman, and MPII. However, because Vision Transformer has a heavy model structure with a large number of parameters and requires a relatively large amount of computation, it costs users a lot to train the model. Accordingly, the based model overcame the insufficient Inductive Bias calculation problem, which requires a large amount of computation by Vision Transformer, with Local Representation through a convolutional neural network structure. Finally, the proposed model obtained a mean average precision of 0.694 on the MS COCO benchmark with 3.28 GFLOPs and 9.72 million parameters, which are 1/5 and 1/9 the number compared to ViTPose, respectively.
참고문헌 (26)
- B. Xiao, "Simple baselines for human pose estimation and tracking", Proceedings of the European Conference on Computer Vision, pp.466~481 (2018)
- 박종혁, "객체 추적 알고리즘을 활용한 딥러닝 기반 실시간 화재 탐지 시스템", 한국컴퓨터정보학회논문지, 27(1), pp.1~8 (2022), 한국컴퓨터정보학회[DOI]
- D. Hwang, "SKU-Net: Improved U-Net using Selective Kernel Convolution for Retinal Vessel Segmentation", Proceedings of The Korea Society of Computer and Information, 26(4), pp.29~37 (2021)
- 양성봉, "CNN 알고리즘의 개선을 통한 대형 이미지에서의 객체식별", 한국컴퓨터정보학회논문지, 25(1), pp.45~53 (2020), 한국컴퓨터정보학회[DOI]
- A. Vaswani, "Attention Is All You Need", Proceedings of the Neural Information Processing Systems (2017)
- A. Dosovitskiy, "An Image Is Worth 16X16Words: Transformers For Image Recognition At Scale", Proceedings of the International Conference on Learning Representations (2021)
- A. . Krizhevsky, "ImageNet Classification with Deep Convolutional Neural Networks", Proceedings of the Neural Information Processing Systems, pp.84~90 (2012)
- C. Szegedy, "Going Deeper with Convolutions", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2015)
- K. He, "Deep Residual Learning for Image Recognition", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.770~778 (2016)
- G. Huang, "Densely Connected Convolutional Networks", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.4700~4708 (2017)
- J. Wang, "Deep High-Resolution Representation Learning for Visual Recognition", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.3349~3364 (2021)
- Y. Xu1, "ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation", Proceedings of the Neural Information Processing Systems (2022)
- T.-Y. Lin, "Microsoft coco:Common objects in context", Proceedings of the European Conference on Computer Vision (2014)
- S. -H. Zhang, "Pose2seg: Detection free human instance segmentation", Proceedings of the IEEE, 2019, 889~898
- M. Andriluka, "2d human pose estimation: New benchmark and state of the art analysis", Proceedings of the Conference on Computer Vision and Pattern Recognition, pp.3686~3693 (2014)
- D. Mehta, "Mobilevit: Light-Weight, General-Purpose, And Mobile-Freindly Vision Transformer", Proceedings of the International Conference on Learning Representations (2022)
- Q. Cheng, "Drone Detection Method Based on MobileViT and CA-PANet", Proceedings of the Electronics, pp.223~239 (2023)
- Y. Yang, "MMViT-Seg: A lightweight transformer and CNN fusion network for COVID-19 segmentation", Proceedings of the Computer Methods and Programs in Biomedicine (2023)
- J. Deng, "Imagenet: A large-scale hierarchical image database", Proceedings of the Conference on Computer Vision and Pattern Recognition, pp.248~255 (2009)
- "Openmmlab pose estimation toolbox and benchmark"[URL]
- S. J. Reddi, "On the convergence of adam and beyond", Proceedings of the International Conference on Learning Representations (2019)
- Z. Yang, "Xlnet: Generalized autoregressive pretraining for language understanding", Proceedings of the Neural Information Processing Systems (2019)