Convergent multi-view geometric error correction with pseudo-inverse projection homography

The paper presents a geometric error correction method for convergent multi-view images, reducing the vertical parallax and non-uniform horizontal disparities over the set of input camera views. The global optimization method using inter-camera geometric relations on an arc is described. Camera orientations and optical centers are adjusted towards a uniform, circular distribution closest to the actual camera setup. Experimental validation shows up to 20% improvements over state-of-the-art w.r.t. reducing geometric errors and reaching better scalability/robustness to large rotation angles between adjacent camera views. The corrected images can be readily used for visualization on autostereoscopic 3D displays, and/or for further processing in an advanced 3D-TV imaging pipeline.

[1]  Zeyu Li,et al.  Camera and projector arrays for immersive 3D video , 2009, IMMERSCOM.

[2]  Vincent Nozick,et al.  Multiple view image rectification , 2011, 2011 1st International Symposium on Access Spaces (ISAS).

[3]  A. Verri,et al.  A compact algorithm for rectification of stereo pairs , 2000 .

[4]  Fei Guo,et al.  Multiview image rectification algorithm for parallel camera arrays , 2014, J. Electronic Imaging.

[5]  Andrea Fusiello,et al.  Quasi-Euclidean uncalibrated epipolar rectification , 2008, 2008 19th International Conference on Pattern Recognition.

[6]  Yo-Sung Ho,et al.  Geometric and colorimetric error compensation for multi-view images , 2014, J. Vis. Commun. Image Represent..

[7]  Peter Eisert,et al.  Extension of epipolar image analysis to circular camera movements , 2003, Proceedings 2003 International Conference on Image Processing (Cat. No.03CH37429).

[8]  C.-C. Jay Kuo,et al.  3D Autostereoscopic Display Image Generation Framework using Direct Light Field Rendering , 2016, 3D Image Processing, Measurement , and Applications.

[9]  Ming C. Lin,et al.  Example-guided physically based modal sound synthesis , 2013, ACM Trans. Graph..

[10]  Yael Pritch,et al.  Scene reconstruction from high spatio-angular resolution light fields , 2013, ACM Trans. Graph..

[11]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

[12]  Richard I. Hartley,et al.  Theory and Practice of Projective Rectification , 1999, International Journal of Computer Vision.

[13]  Qionghai Dai,et al.  Light field depth estimation exploiting linear structure in EPI , 2015, 2015 IEEE International Conference on Multimedia & Expo Workshops (ICMEW).

[14]  Rajiv Gupta,et al.  Computing matched-epipolar projections , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[15]  Yo-Sung Ho,et al.  Geometric error correction of convergent multi-view images , 2015 .

[16]  Vincent Nozick,et al.  Epipolar Rectification for Autostereoscopic Camera Setup , 2010 .

[17]  Robert Laganière,et al.  Projective Rectification Of Image Triplets From The Fundamental Matrix , 2006, 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings.

[18]  Yo-Sung Ho,et al.  An efficient image rectification method for parallel multi-camera arrangement , 2011, IEEE Transactions on Consumer Electronics.

[19]  Zhengyou Zhang,et al.  A Flexible New Technique for Camera Calibration , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[20]  An Ping,et al.  A rectification algorithm for un-calibrated multi-view images based on SIFT features , 2010, 2010 International Conference on Audio, Language and Image Processing.

[21]  Paul F. Whelan,et al.  Projective rectification from the fundamental matrix , 2005, Image Vis. Comput..