人體三維姿態常利用監督式學習的方法來預估，這種方法的缺點是需要大量的二維骨架、三維骨架配對資料。但收集三維骨架資料的成本相當昂貴且花時間，因此近年來許多研究開始使用弱監督式學習的方法在預估人體三維姿態，希望能只透過少量的3D data就能有效的訓練。
在弱監督式學習的方法中，有人提出分別考慮時間資訊或多視角資訊來訓練模型。在本論文中，我們提出一個基於對抗式生成網路的弱監督式學習方法，同時整合時間資訊跟多視角資訊，取得目前為止以弱監督式學習為基礎的方法中最好的人體姿態預估的結果。

Three-dimensional human pose estimation is usually conducted in a supervised manner. Developing a supervised model requires a large amount of 2D and 3D paired skeletons. Collecting labeled 3D skeletons, however, is quite expensive and time-consuming. Many studies thus have been proposed to estimate 3D human pose in a weakly supervised manner. In this way, only a small amount of 3D data is needed to train an effective model.
In weakly supervised learning methods, some studies have been independently proposed to consider time information or multiview information for training. In this thesis, we propose a weakly supervised learning method based on the generative adversarial network. We jointly consider time information and multiview information and achieve state-of-the-art results.

摘要 i
Abstract ii
List of Figures iii
List of Tables iv
Contents v
1 Introduction 1
1.1 Motivation 1
1.2 System Overview 3
1.3 Contributions 4
1.4 Thesis Organization 4
2 Related Works 6
2.1 Traditional Methods 6
2.2 Deep-based Methods 6
2.3 Short Summary 7
3 Proposed Methods 8
3.1 RepNet 8
3.2 Considering Temporal Information 9
3.3 Considering Multiview Information 11
3.4 Short Summary 12
4 Evaluation 13
4.1 Dataset and Experimental Settings 13
4.2 Performance 16
4.2.1 Comparing with Supervised Methods 16
4.2.2 Comparing with Weakly-Supervised Methods 17
5 Conclusion and Future Works 22
5.1 Conclusion 22
5.2 Future Works 23
References 24
Appendix 27

[1] Chen, X., Lin, K.Y., Liu, W., Qian, C., Lin, L.:Weakly-supervised discovery of geometry-aware representation for 3d human pose estimation. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2019)
[2] Ionescu, C., Papava, D., Olaru, V., Sminchisescu, C.: Human3.6m: Large scale datasets and predictive methods for 3d human sensing in natural environments. IEEE Transactions on Pattern Analysis and Machine Intelligence 36(7), 1325--1339 (2014)
[3] Li, Z.,Wang, X.,Wang, F., Jiang, P.: On boosting single-frame 3d human pose estimation via monocular videos. In Proceedings of International Conference on Computer Vision (2019)
[4] Martinez, J., Hossain, R., Romero, J., Little, J.J.: A simple yet efective baseline for 3d human pose estimation. In Proceedings of IEEE International Conference on Computer Vision (2017)
[5] Park, S., Hwang, J., Kwak, N.: 3d human pose estimation using convolutional neural networks with 2d pose information. In Proceedings of European Conference on Computer Vision Workshops, pp. 156--169 (2016)
[6] Pavllo, D., Feichtenhofer, C., Grangier, D., Auli, M.: 3d human pose estimation in video with temporal convolutions and semi-supervised training. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2019)
[7] Rhodin, H., Salzmann, M., Fua, P.: Unsupervised geometry-aware representation for 3d human pose estimation. In Proceedings of European Conference on Computer Vision (2018)
[8] Rogez, G., Weinzaepfel, P., Schmid, C.: Lcr-net++: Multi-person 2d and 3d pose detection in natural images. IEEE Transactions on Pattern Analysis and Machine Intelligence 42(5), 1146--1161 (2020)
[9] Wandt, B., Rosenhahn, B.: Repnet weakly supervised training of an adversarial reprojection network for 3d human pose estimation. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2019)
[10] Newell, A., Yang, K, Deng, J.: Stacked hourglass networks for human pose estimation. In Proceedings of European Conference on Computer Vision (2016)
[11] Lee, H. J., Chen, Z.: Determination of 3D human body postures from a single view. Computer Vision, Graphics and Image Processing, 30:148–-168 (1985)
[12] Jiang, H.: 3d human pose reconstruction using millions of exemplars. In Proceedings of International Conference on Pattern Recognition, 1674–1677 (2010)
[13] Akhter, I., Black, M. J.: Pose-conditioned joint angle limits for 3D human pose reconstruction. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2015)
[14] Bogo, F., Kanazawa, A., Lassner, C., Gehler, P., Romero, J., Black, M. J.: Keep it SMPL: Automatic estimation of 3d human pose and shape from a single image. In Proceedings of European Conference on Computer Vision (2016)
[15] Ramakrishna, V., Kanade, T., Sheikh, Y.: Reconstructing 3D Human Pose from 2D Image Landmarks. In Proceedings of European Conference on Computer Vision (2012)
[16] Wang, C., Wang, Y., Lin, Z., Yuille, A. L., Gao, W.: Robust estimation of 3d human poses from a single image. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2014)
[17] Zhou, X., Zhu, M., Leonardos, S., Daniilidis, K.: Sparse representation for 3d shape estimation: A convex relaxation approach. IEEE Transactions on Pattern Analysis and Machine Intelligence, 99(1) (2016)
[18] Zhou, X., Zhu, M., Leonardos, S., Derpanis, K. G., Daniilidis, K.: Sparseness meets deepness: 3d human pose estimation from monocular video. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2016)
[19] Agarwal, A., Triggs, B.: 3D human pose from silhouettes by relevance vector regression. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2004)
[20] Mori, G., Malik, J.: Recovering 3D human body configurations using shape contexts. IEEE Transactions on Pattern Analysis and Machine Intelligence, 28(7):1052–1062 (2006)
[21] Bo, L. F., Sminchisescu, C., Kanaujia, A., Metaxas, D. N.: Fast algorithms for large scale conditional 3D prediction. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2008)
[22] Shakhnarovich, G., Viola, P. A., Darrell, T. J.: Fast pose estimation with parameter-sensitive hashing. In Proceedings of IEEE International Conference on Computer Vision (2003)
[23] Pavlakos, G., Zhou, X., Derpanis, K. G., Daniilidis, K.: Coarse-to-fine volumetric prediction for single-image 3D human pose. In Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition (2017)
[24] Zhou, X., Sun, X., Zhang, W., Liang, S., Wei, Y.: Deep kinematic pose regression. In Proceedings of European Conference on Computer Vision Workshops (2016)
[25] Andriluka, M., Pishchulin, L., Gehler, P., Schiele, B.: 2D Human Pose Estimation: New Benchmark and State of the Art Analysis. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (2014)
[26] Hurley, J. R., Cattell, R. B.: The procrustes program: Producing direct rotation to test a hypothesized factor structure. Behavioral Science (1962)