High-Speed Periodic Motion Reconstruction Using an Off-the-shelf Camera with Compensation for Rolling Shutter Effect Jeong-Jik Seo, Wissam J. Baddar, Hyung-Il Kim, and Yong Man Ro (B ) Image and Video Systems Lab, School of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea {jj.seo,wisam.baddar,hyungil.kim,ymro}@kaist.ac.kr Abstract. In recent years, high-speed signal reconstruction with sub- Nyquist sampling have attracted the attention of researchers in the signal processing field. Nonetheless, such methods have been limited either by the need to utilize multiple cameras, or relying on newly designed imag- ing hardware. In this paper, we propose a high-speed periodic motion reconstruction method, obtained by randomly delaying the camera expo- sure. This allows it to utilize a conventional off-the-shelf camera. In addition, the proposed method compensates the rolling shutter effect, which is inevitable if the camera’s image sensor is made of complemen- tary metal-oxide semiconductor (CMOS), while reconstructing the high- speed periodic motion. Exhaustive and comparative experiments have been conducted to validate the proposed method, which showed promis- ing performance in terms of reconstruction error, and effective compen- sation of the rolling shutter effect. Keywords: Computational imaging · Compressive sensing · High speed motion · Sparse reconstruction · Rolling shutter effect 1 Introduction According to the Nyquist sampling theorem, the sampling rate should be at least twice the bandwidth of a signal in order to avoid information loss (i.e., aliasing). Therefore, high-speed motions which can be observed in real-life (e.g., hand mix- ers or electric fan motions rotating over 60 rounds per second) cannot be captured by a conventional low-frame rate cameras (e.g., 30 frames per second (fps)). In other words, in order to capture such fast motions, a high-speed camera which can cover high sampling rate (with at least over 120 fps) is required. However, the high-speed cameras have been known to be relatively expensive [7]. In addi- tion, the short aperture time of high-speed cameras induces light-inefficiency [5]. Recently, in image processing area, there have been various research efforts [4, 6– 8] concerning the reconstruction of high-speed motion from the low-frame rate cameras. In [4, 8], the authors proposed reconstructing the high-speed video by c  Springer International Publishing Switzerland 2015 Y.-S. Ho et al. (Eds.): PCM 2015, Part II, LNCS 9315, pp. 311–320, 2015. DOI: 10.1007/978-3-319-24078-7 31