Abstract
A selective vision system sequentially collects evidence to answer a specific question with a desired level of confidence. Efficiency comes from processing the scene only where necessary, to the level of detail necessary, and with only the necessary operators. Knowledge representation and sequential decision making are central issues for selective vision, which takes advantage of prior knowledge of a domain's abstract and geometrical structure (e.g., “part-of” and “adjacent” relationships), and also uses information from a scene instance gathered during analysis. The TEA-1 selective vision system uses Bayes nets for representation, benefit-cost analysis for control of visual and nonvisual actions; and its data structures and decision-making algorithms provide a general, reusable framework. TEA-1 solves the T-world problem, an abstraction of a large set of scene domains and tasks. Some factors that affect the success of selective perception are analyzed by using TEA-1 to solve ensembles of randomly produced, simulated T-world problems. Experimental results with a real-world T-world problem, dinner table scenes, are also presented.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Agosta, J.M. 1991. Probabilistic Recognition Networks, An Application of Influence Diagrams to Visual Recognition, Ph.D. thesis, Stanford University.
Anderson, S.K., Olesen, K.G., Jensen, F.V., and Jensen, F. 1989. HUGIN—A shell for building belief universes for expert systems,Proc. 11th Intern. Joint Conf. Artif. Intell., Detroit, pp. 1080–1085.
Ballard, D.H., and Brown, C.M. 1992. Principles of animate vision,Comput. Vis., Graphics, Image Process. Image Understanding, 56(1):3–21.
Bolle, R.M., Califano, A., and Kjeldsen, R. 1990. Data and model driven foveation,Proc. IEEE Intern. Conf. Patt. Recog., Atlantic City, pp. 1–7.
Bolles, R.C. 1977. Verification vision for programmable assembly,Proc. 5th Intern. Joint Conf. Artif. Intell. Cambridge, MA, pp. 569–575.
Burt, P.J. 1988. Smart sensing within a pyramid vision machine,Proc. IEEE, 76(8): 1006–1015.
Charniak, E. 1991. Bayesian networks without tears,AI Magazine, 12(4):50–63.
Chen, C., and Mulgaonkar, P.G. 1992. Automatic vision programming,Comput. Vis., Graphics, Image Process. Image Understanding, 55(2):170–183.
Chou, P.B., and Brown, C.M. 1990. The theory and practice of Bayesian image labeling,Intern. J, Comput. Vis., 4(3): 185–210.
Clark, J.J., and Ferrier, NJ. 1988. Modal control of an attentive vision system,Proc. 2nd Intern. Conf. Comput. Vis. Tampa, FL, pp. 514–523.
Clemen, R.T. 1991.Making Hard Decisions: An Introduction to Decision Analysis. PWS-Kent Publishing: Boston.
Cooper, G. 1990. The computational complexity of probabilistic inference using belief networks,Artificial Intelligence, 42(2–3): 393–405.
Dean, T., Camus, T., and Kirman, J. 1990. Sequential decision making for active perception,Proc. DARPA Image Understanding Workshop, Pittsburgh, pp. 889–894.
Dean, T.L., and Wellman, M.P. 1991.Planning and Control, Morgan Kaufmann: Los Altos, CA.
Durrant-Whyte, H.F. 1988.Integration, Coordination, and Control of Multi-Sensor Robot Systems, Kluwer Academic: Norwell, MA.
Elfes, A. 1992. Dynamic control of robot perception using multi-property inference grids,Proc. IEEE Intern. Conf. Robot. Autom., pp. 2561–2567.
Feldman, J., and Sproull, R. 1977. Decision theory and artificial intelligence II: The hungry monkey,Cognitive Science, 1:158–192.
Garvey, T. 1976. Perceptual strategies for purposive vision, Tech. Rept. 117, SRI AI Center.
Hager, G.D. 1990.Task-Directed Sensor Fusion and Planning: A Computational Approach, Kluwer Academic: Norwell, MA.
Heckerman, D., Horvitz, E., and Middleton, B. 1993. An approximate nonmyopic computation for value of information,IEEE Trans. Patt. Anal. Mach. Intell., 15(3):292–298.
Henrion, M. 1990. An introduction to algorithms for inference in belief nets,Uncertainty in AI 5, North-Holland: New York, pp. 129–138.
Hutchinson, S.A., and Kak, A.C. 1989. Planning sensing strategies in a robot work cell with multi-sensor capabilities,”IEEE J. Robot. Autom., 5(6):765–783.
Jensen, F.V., Christensen, H.I., and Nielsen, J. 1992. Bayesian methods for interpretation and control in multi-agent vision systems,Proc. Appl. AI X: Machine Vision and Robotics.
Jensen, F.V., Lauritzen, S.L., and Olesen, K.G. 1990. Bayesian updating in recursive graphical models by local computations,Computat. Stat. Quart., 4:269–282.
Krotkov, E.P. 1989.Active Computer Vision by Cooperative Focus and Stereo, Springer-Verlag: New York.
Lauritzen, S.L., and Spiegelhalter, D.J. 1988. Local computations with probabilities on graphical structures and their application to expert systems,J. Roy. Stat. Soc. B, 50(2): 157–224.
Levitt, T., Binford, T., Ettinger, G., and Gelband, P. 1989. Probability-based control for computer vision,Proc. DARPA Image Understanding Workshop, Palo Alto, pp. 355–369.
Levitt, T.S. 1986. Model-based probabilistic inference in hierarchical hypothesis spaces,Uncertainty in AI, North-Holland: New York, pp. 347–356.
Mann, W.B., and Binford, T.O. 1992. An example of 3D interpretation of images using Bayesian networks,Proc. DARPA Image Understanding Workshop, pp. 793–801.
Pearl, J. 1986. Fusion, propagation, and structuring in Bayesian networks,Artificial Intelligence, 29(3):241–288.
Pearl, J. 1988.Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference, Morgan Kaufmann: Los Altos, CA.
Peot, M.A. and Shachter, R.D. 1991. Fusion and propagation with multiple observations in belief networks,Artificial Intelligence, 48(3): 299–318.
Reece, D.A. and Shafer, S.A. 1992. Planning for perception in robot driving,Proc. DARPA Image Understanding Workshop, pp. 953–960.
Rimey, R.D. 1993. Control of selective perception using Bayes nets and decision theory, Ph.D. thesis, Department of Computer Science, University of Rochester. (Also available as Tech. Rept. 468, Department of Computer Science, University of Rochester, 1993.)
Rimey, R.D., and Brown, C.M. 1991. Controlling eye movements with hidden Markov models,Intern. J. Comput. Vis., 7(1):47–65.
Rimey, R.D., and Brown, C.M. 1992. Task-oriented vision with multiple Bayes nets. In A. Blake and A. Yuille, eds.,Active Vision, MIT Press: Cambridge, MA, pp. 217–236.
Sarkar, S. and Boyer, K.L. 1993. Integration, inference, and management of spatial information using Bayesian networks: Perceptual organization,IEEE Trans. Patt. Anal. Mach. Intell, 15(3): 256–272.
Shachter, R.D. 1986. Evaluating influence diagrams,Operations Research, 34(6):871–882.
Shafer, G., and Pearl, J., eds. 1990.Readings in Uncertain Reasoning, Morgan Kaufmann: Los Altos, CA.
Tarabanis, K., Tsai, R.Y., and Allen, P.K. 1991. Automated sensor planning for robotic vision tasks,Proc. IEEE Intern. Conf. Robot. Autom., pp. 76–82.
Tsotsos, J. 1989. The complexity of perceptual search tasks,Proc. 11th Intern. Joint Conf. Artif. Intell., Detroit, pp. 1571–1577.
von Kaenel, P.A., Brown, C.M., and Rimey, R.D. 1993. Goal-oriented dynamic vision, Tech. Rept. 469, Department of Computer Science, University of Rochester.
Wixson, L., and Ballard, D. 1994. Using intermediate objects to improve the efficiency of visual search,Intern. J. Comput. Vis., this issue.
Wu, H.L., and Cameron, A. 1990. A Bayesian decision theoretic approach for adaptive goal-directed sensing,Proc. 3rd Intern. Conf. Comput. Vis., Osaka, pp. 563–567.
Yarbus, A. 1967.Eye Movements and Vision, Plenum Press: New York.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Rimey, R.D., Brown, C.M. Control of selective perception using bayes nets and decision theory. Int J Comput Vision 12, 173–207 (1994). https://doi.org/10.1007/BF01421202
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01421202