Motion - Stereo Integration for Depth Estimation
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part II
| Hi-index | 0.00 |
In this paper, a new hierarchical stereo algorithm is presented. The algorithm matches individual pixels in corresponding scanlines by minimizing a cost function. Several cost functions are compared. The algorithm achieves atremendous gain in speed and memory requirements by implementing it hierarchically. The images are downsampledan optimal number of times and the disparity map of a lowerlevel is used as 'offset' disparity map at a higher level. Animportant contribution consists of the complexity analysisof the algorithm. It is shown that this complexity is independent of the disparityrange. This result is also used todetermine the optimal number of downsample levels. Thisspeed gain results in the ability to use more complex (compute intensive) cost functions that deliver high quality disparity maps. Another advantage of this algorithm is thatcost functions can be chosen independent of the optimisation algorithm. Finally, the algorithm was carefully implemented so that a minimal amount of memory is used. It hasproven its efficiency on large images with a high disparityrange as well as its quality. Examples are be given in thispaper.