ABSTRACT
Motion segmentation is about separating moving objects from the background in video. If the motion is captured by a stationary camera, the solution is trivial. However, if the motion is captured by a moving camera, then the problem is hard to solve. Algorithms proposed so far usually assume the motion of the scene and the camera are both unknown. Consequently, such algorithms are generally computationally expensive and can be fragile for real-time vision applications. In this work, we propose to simplify the problem by using inertial sensors to measure the camera motion directly. The problem can then be simplified considerably, requiring only a simple line fitting algorithm in order to discriminate between the background and the moving objects. We believe the move towards incorporating inertial sensors to vision applications is essential in supporting robust real-time vision applications in future.
- Weber, J., and Malik, J., 1997, Rigid body segmentation and shape description from dense optical flow under weak perspective, Pattern Analysis and Machine Intelligence, IEEE Trans., Vol. 19, Issue 2, Feb., pp. 139--143. Google ScholarDigital Library
- Weiss, Y., 1997, Smoothness in layers: Motion segmentation using nonparametric mixture estimation, Conf. on Computer Vision and Pattern Recognition. Proceedings, pp. 520--526. Google ScholarDigital Library
- Ogale, A. S., Fermuller, C., and Aloimonos, Y., 2005. Motion segmentation using occlusions, Pattern Analysis and Machine Intelligence, IEEE Trans., Vol. 27, Issue 6, Jun 2 pp. 988--992. Google ScholarDigital Library
- Adiv, G., 1985. Determining three-dimensional motion and structure from optical flow generated by several moving objects, Pattern Analysis and Machine Intelligence, IEEE Trans., Vol. 7, Issue 4, July, pp. 384--401.Google ScholarDigital Library
- Zhang, Z, Faugeras, O. D., and Ayache, N., 1988. Analysis Of A Sequence Of Stereo Scenes Containing Multiple Moving Objects Using Rigidity Constraints, Computer Vision. Second Int Conf on, pp. 177--186Google Scholar
- Wills, J., Agarwal, S., and Belongie, S., 2003. What went where, Computer Vision and Pattern Recognition. Proceedings. 2003 IEEE Computer Society Conference on, Vol. 1, pp. I--37 - I--44 vol.1. Google ScholarDigital Library
- Mansouri, A.-R., and Konrad, J., 2003. Multiple motion segmentation with level sets, Image Processing, IEEE Trans., Vol. 12, Issue 2, Feb. pp. 201--220. Google ScholarDigital Library
- Tong, W. S, Tang, C. K., and Medioni, G., 2004. Simultaneous two-view epipolar geometry estimation and motion segmentation by 4D tensor voting, Pattern Analysis and Machine Intelligence, IEEE Trans., Vol. 26, Issue 9, Sept. pp. 1167--1184. Google ScholarDigital Library
- Torr, P. H. S., Szeliski, R., and Anandan, P., 2001. An integrated Bayesian approach to layer extraction from image sequences, Pattern Analysis and Machine Intelligence, IEEE Trans., Vol. 23, Issue 3, March, pp. 297--303. Google ScholarDigital Library
- Bober, M., and Kittler, J., 1994. Robust motion analysis, Computer Vision and Pattern Recognition. Proceedings. 1994 IEEE Computer Society Conference on, pp. 947--952Google Scholar
- Fan. Z, Zhou. J, and Wu, Y, 2004. Multibody motion segmentation based on simulated annealing", Computer Vision and Pattern Recognition. Proceedings. 2004 IEEE Computer Society Conference on. Vol. 1, pp. I--776 - I--781.Google Scholar
- Vieville, T., 1997. A few steps towards 3D active vision, Springer-Verlag. Google ScholarDigital Library
- Lobo, J., and Dias, J., 2003. Vision and inertial sensor cooperation using gravity as a vertical reference, Pattern Analysis and Machine Intelligence, IEEE Trans., Vol. 25, Issue 12, Dec. pp. 1597--1608. Google ScholarDigital Library
- Hu, Z. and Uchimura, K., 2004, Real-time data fusion on tracking camera pose for direct visual guidance, Intelligent Vehicles Symposium, IEEE, pp. 842--847.Google Scholar
- Bhanu, B., Roberts, B., and Ming, J., 1990. Inertial navigation sensor integrated motion analysis for obstacle detection", Robotics and Automation, IEEE International Conf., pp. 954--959 vol.2.Google Scholar
- Tao, Y, Hu H, and Zhou, H, 2005. Integration of vision and inertial sensors for home-based rehabilitation, InerVis2005, 2nd Workshop on Integration of Vision and Inertial Sensors, IEEE Int. Conf. on Robotics and Automation.Google Scholar
- Chroust, S. G. and Vincze, M., 2004. Fusion of vision and inertial data for motion and structure estimation, J. of Robotic Systems 21(2), pp. 73--83. Google ScholarDigital Library
- Qian, G., Zheng, Q., and Chellappa, R., 2000. Reduction of inherent ambiguities in structure from motion problem using inertial data, Image Processing, Int Conf., Vol. 1, pp. 204--207 vol.1.Google Scholar
Index Terms
- Motion segmentation using inertial sensors
Recommendations
Demonstration abstract: upper body motion capture system using inertial sensors
IPSN '14: Proceedings of the 13th international symposium on Information processing in sensor networksMotion capture plays an important role in interactive gaming, animation, film industry and navigation. The existing camera-based motion capture studios are expensive and require a clear line of sight; hence they cannot be applied to ubiquitous ...
Automated human motion segmentation via motion regularities
Analysis and reuse of human motion capture (mocap) data play an important role in animation, games and medical rehabilitation. In various mocap-based animation techniques, motion segmentation is regarded as one of the fundamental functions. Many ...
Efficient Dense Rigid-Body Motion Segmentation and Estimation in RGB-D Video
Motion is a fundamental grouping cue in video. Many current approaches to motion segmentation in monocular or stereo image sequences rely on sparse interest points or are dense but computationally demanding. We propose an efficient expectation---...
Comments