Dense Optic Flow with a Bayesian Occlusion Model
This paper presents a dense optic flow algorithm based on quite simple local probability assumptions. Due to the explicit derivation of the correspondence concept in a probability-theoretical framework, occlusion probability evolves straight-forwardly from the model for each pixel. Initialized with a similarity measure based on single pixels, an iterated diffusion step propagates local information across the image, while occlusion probability is used to inhibit flow information transfer across depth discontinuities, which prevents flow smoothing at 3d object boundaries. The inhibition is thereby not artificially modelled by some heuristically chosen parameters, but arises directly from the Bayesian correspondence model. The algorithm structure can be interpreted as a recurrent neural network, where matched points have reached a stable state, while others (e.g. those in homogeneous areas) keep receiving information from regions more and more far away until they converge, this way overcoming the aperture problem. The massive parallel structure allows for and demands a real hardware implementation of the system.
KeywordsTest Patch Neighboring Pixel Foreground Object Stereo Match Depth Discontinuity
Unable to display preview. Download preview PDF.
- 1.Kanade, T., Okutomi, M.: A stereo matching algorithm with an adaptive window: Theory and experiment. IEEE Transactions on PAMI 16, 920–932 (1994)Google Scholar
- 2.Horn, B., Schunck, B.: Determining optical flow. AI 17, 185–204 (1981)Google Scholar
- 3.Marroquin, J., Velasco, F., Rivera, M., Nakamura, M.: Gauss-markov measure field models for low-level vision. IEEE Transactions on PAMI 23, 337–348 (2001)Google Scholar
- 5.Zitnick, C., Kanade, T.: A cooperative algorithm for stereo matching and occlusion detection. IEEE Transactions on PAMI 22, 675–684 (2000)Google Scholar
- 6.Vasconcelos, N., Lippman, A.: Empirical bayesian motion segmentation. IEEE Transactions on PAMI 23, 217–221 (2001)Google Scholar
- 7.Torr, P., Szeliski, R., Anandan, P.: An integrated bayesian approach to layer extraction from image sequences. IEEE Transactions on PAMI 23, 297–303 (2001)Google Scholar
- 9.Konrad, J., Dubois, E.: Bayesian estimation of motion vector fields. IEEE Transactions on PAMI 14, 910–927 (1992)Google Scholar
- 11.Memin, E., Perez, P.: Dense estimation and object-based segmentation of the optical flow with robust techniques. IEEE Transactions on IP 7, 703–719 (1998)Google Scholar
- 12.Perwass, C., Sommer, G.: An iterative bayesian technique for dense image point matching. In: Proceedings of Dynamic Perception, pp. 283–288 (2002)Google Scholar
- 13.Egnal, G., Wildes, R.: Detecting binocular half-occlusions: Empirical comparisons of four approaches. In: IEEE Conference CVPR, pp. 466–473 (2000)Google Scholar
- 14.Koeser, K.: Dense image point matching with explicit occlusion detection for stereo disparity and optic flow. Master’s thesis, CAU Kiel (2003), available at http://www.ks.informatik.uni-kiel.de
- 15.Galvin, B., McCane, B., Novins, K., Mason, D., Mills, S.: Recovering motion fields: An evaluation of eight optical flow algorithms. In: Proc. of BMVC, pp. 195–204 (1998)Google Scholar