Abstract
Most computer vision algorithms have parameters that must be hand-selected using expert knowledge. These parameters hinder the use of large computer vision systems in real-world applications. In this work, a method is presented for automatically and continuously tuning the parameters of algorithms in a real-time modular vision system. In the training phase, a human expert teaches the system how to adapt the algorithm parameters based on training data. During operation, the system measures features from the inputs and outputs of each module and decides how to modify the parameters. Rather than learning good parameter values in absolute terms, incremental changes are modelled based on relationships between algorithm inputs and outputs. These increments are continuously applied online so that parameters stabilise to suitable values. The method is demonstrated on a three-module people-tracking system for video surveillance.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Cai, X.C., Sowmya, A., Trinder, J.: Learning parameter tuning for object extraction. In: Asian Conf. on Computer Vision, pp. 868–877 (2006)
Crowley, J.L., Hall, D., Emonet, R.: Autonomic computer vision systems. In: Blanc-Talon (ed.) Advanced Concepts for Intell. Vision Systems (August 2007)
Crowley, J.L., Hall, D., Emonet, R.: An automatic approach for parameter selection in self-adaptive tracking. In: VISAPP, Portugal. Springer, Heidelberg (February 2006)
Ferryman, J., Shahrokni, A.: An overview of the PETS 2009 challenge. In: PETS Workshop, Miami, USA, pp. 25–30. IEEE Computer Society, Los Alamitos (June 2009)
Scotti, G., Marcenaro, L., Regazzoni, C.S.: SOM based algorithm for video surveillance system parameter optimal selection. In: AVSS, pp. 370–375 (2003)
Hall, D.: Automatic parameter regulation for a tracking system with an auto-critical function. In: CAMP Workshop, pp. 39–45 (July 2005)
Hall, D.: Automatic parameter regulation of perceptual systems. Image and Vision Computing 24(8), 870–881 (2006)
Huang, Y., Brown, M.S.: User-assisted ink-bleed correction for handwritten documents. In: Joint Conf. on Digital Libraries, pp. 263–271. ACM, New York (2008)
Marsico, M., Nappi, M., Riccio, D.: Multibiometric people identification: A self-tuning architecture. In: Int’l Conf. on Advances in Biometrics, pp. 980–989. Springer, Heidelberg (2009)
De Marsico, M., Nappi, M., Riccio, D.: A self-tuning people identification system from split face components. In: Wada, T., Huang, F., Lin, S. (eds.) PSIVT 2009. LNCS, vol. 5414, pp. 1–12. Springer, Heidelberg (2009)
Murino, V., Foresti, G.L., Regazzoni, C.S.: A distributed probabilistic system for adaptive regulation of image-processing parameters. IEEE Trans. Systems, Man and Cybernetics 26(1), 1–20 (1996)
Ramesh, V., Haralick, R.M., Zhang, X., Nadadur, D.C., Thornton, K.B.: Automatic selection of tuning parameters for feature extraction sequences. In: CVPR, pp. 672–677 (1994)
Robertson, P., Brady, M.: Adaptive image analysis for aerial surveillance. IEEE Intelligent Systems 14, 30–36 (1999)
Shekhar, C., Burlina, P., Moisan, S.: Design of self-tuning IU systems. In: Image Understanding Workshop, pp. 529–536 (1997)
Shekhar, C., Moisan, S., Vincent, R., Burlina, P., Chellappa, R.: Knowledge-based control of vision systems. Image and Vision Computing 17(9), 667–683 (1999)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Sherrah, J. (2010). Learning to Adapt: A Method for Automatic Tuning of Algorithm Parameters. In: Blanc-Talon, J., Bone, D., Philips, W., Popescu, D., Scheunders, P. (eds) Advanced Concepts for Intelligent Vision Systems. ACIVS 2010. Lecture Notes in Computer Science, vol 6474. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-17688-3_39
Download citation
DOI: https://doi.org/10.1007/978-3-642-17688-3_39
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-17687-6
Online ISBN: 978-3-642-17688-3
eBook Packages: Computer ScienceComputer Science (R0)