Action Recognition Model to Monitor Illegal Dumping using Zoom-In Image Sequence Data
0
조회
0
다운로드
2
인용
0
참고문헌
키워드
초록
In this study, we propose an action recognition model that provides generalized performance regardless of camera location and distance between the camera and human. The proposed model consists of two-stage networks, namely, human detection and action recognition. The proposed method operates on video frames that are resized by a new zoom- in method using pretrained Yolo v3. To use temporal information, which is regarded as a critical factor in action recognition, we adopt the R(2+1)D model, which is a factorized model capable of representing more complex networks. The proposed Zoom-In method yields generalized performance regardless of distance. In an experiment, the proposed method exhibited accuracies of 96.07%, 96.61%, and 94.55% in the short, medium, and long ranges in which our datasets were employed, respectively.
Abstract
In this study, we propose an action recognition model that provides generalized performance regardless of camera location and distance between the camera and human. The proposed model consists of two-stage networks, namely, human detection and action recognition. The proposed method operates on video frames that are resized by a new zoom- in method using pretrained Yolo v3. To use temporal information, which is regarded as a critical factor in action recognition, we adopt the R(2+1)D model, which is a factorized model capable of representing more complex networks. The proposed Zoom-In method yields generalized performance regardless of distance. In an experiment, the proposed method exhibited accuracies of 96.07%, 96.61%, and 94.55% in the short, medium, and long ranges in which our datasets were employed, respectively.
참고문헌 (18)
- Al-Qizwini M, "Deep learning algorithm for autonomous driving using googlenet", Ieee Int Veh Sym, pp.89~89 (2017)
- Yan K, "Face recognition based on convolution neural network", IEEE 36th Chinese Control Conference (CCC), pp.4077~4081 (2017)
- Viitaniemi V, "Techniques for image classification, object detection and object segmentation", International Conference on Advances in Visual Information Systems, pp.231~234 (2008)
- Lima T, "Human action recognition with 3d convolutional neural network", IEEE Latin American Conference on Computational Intelligence (LA-CCI), pp.1725~1732 (2017)
- Karpathy A, "Large-scale video classification with convolutional neural networks", Proc Cvpr Ieee, pp.1725~1732 (2014)
- 윤기민, "Vision‐based garbage dumping action detection for real‐world surveillance platform", ETRI Journal, 41(4), pp.494~505 (2019), 한국전자통신연구원[DOI]
- Mahankali S, "Identification of illegal garbage dumping with video analytics", IEEE International Conference on Advances in Computing, Communications and Informatics (ICACCI), pp.2403~2407 (2018)
- Gholamrezaii M, "Human activity recognition using 2d convolutional neural networks", IEEE Iranian Conference on Electrical Engineering (ICEE), pp.1682~1686 (2019)
- Hoang VD, "Action recognition based on sequential 2d-cnn for surveillance systems", In IECON 2018-44th Annual Conference of the IEEE Industrial Electronics Society, pp.3225~3230 (2018)
- Tran D, "A closer look at spatiotemporal convolutions for action recognition", Proc Cvpr Ieee, pp.6450~6459 (2018)
- Ren S, "Towards real-time object detection with region proposal networks", Advances in Neural Information Processing Systems, 28, pp.91~99 (2015)
- Redmon J, "You only look once: Unified, real-time object detection", Proc Cvpr Ieee, pp.779~788 (2019)
- Jiao L, "A survey of deep learning-based object detection", IEEE Access, 7, pp.128837~128868 (2019)
- Redmon J, "Yolov3 : An incremental improvement", arXiv preprint arXiv:1804.02767, ~ (2018)
- Agarap AF, "Deep learning using rectified linear units(relu)", arXiv preprint arXiv:1803.08375, ~ (2018)
- Lin TY, "Microsoft coco: Common objects in context", Proc Cvpr Ieee, pp.740~755 (2014)
- He K, "Deep residual learning for image recognition", Proc Cvpr Ieee, pp.770~778 (2016)
- Paszke A, "Pytorch: An imperative style, high-performance deep learning library", Advances in Neural Information Processing Systems, 32, pp.8026~8037 (2019)