The hybrid recursive matching algorithm (HRM) is a software solution that runs in real-time on standard PC platforms for high-resolution TV formats (SDTV and higher). It unites the advantages of block-recursive disparity matching and pixel-recursive optical flow estimation in one common scheme.
The main idea is to use three neighbouring spatio-temporal candidates as input for the block-recursive disparity estimation. The rationale is that such candidate vectors are the most likely to provide a good estimate of the disparity for the current pixel. In addition, a further update vector is tested against the best candidate. This update vector is computed by applying a local, pixel-recursive process to the current block using the selected candidate of block-recursion as a start vector. The final vector is obtained by comparing the update vector from the pixel recursive stage with the selected vector from the block-recursive one. As usual, a normalized cross correlation is used as matching and selection criterion.
The HRM scheme has two main advantages in comparison to common approaches. On the one hand, the recursive structure speeds up the analysis process dramatically and allows real-time processing therefore. On the other hand, due to an efficient strategy of testing particular vector candidates, the combined choice of spatial and temporal candidates yields disparity vectors of relatively high reliability and high spatio-temporal smoothness. The latter aspect is important to reduce mismatches in critical regions with homogeneous, periodic or similar textures.
Competencies in Stereo Matching
- stereo and multi-view analysis for real-time applications
- advanced stereo by exploiting colour segmentation, motion detection and trifocal costraints
Publications
R. Tanger, N. Atzpadin, M. Müller, C. Fehn, P. Kauff, C. Herpel, "Depth Acquisition for Post-Production Using Trinocular Camera Systems and Trifocal Constraint", Proc, of Int. Broadcast Conference (IBC 2006), Amsterdam, The Netherlands, September 2006.
N. Atzpadin, P. Kauff, O. Schreer, "Stereo Analysis by Hybrid Recursive Matching for Real-Time Immersive Video Stereo Analysis by Hybrid Recursive Matching for Real-Time Immersive Video Conferencing", IEEE Trans. on Circuits and Systems for Video Technology, Special Issueon Immersive Telecommunications, pp. 321-334, Vol. 14, No. 3, January 2004.