Summary
The ability to predict the fundamental performance of model-based object recognition is essential for transforming the object recognition field from an art to a science, and to speed up the design process for recognition systems. In this chapter, we address the performance—prediction problem in the context of a common recognition task, where both model objects and scene data are represented by locations of 2D point features. The criterion used for estimating matching quality is based on the number of consistent data/model feature pairs, which we refer to as “votes.” We present a theoretical framework for prediction of lower and upper bounds on the probability of correctly recognizing model objects from scene data. The proposed framework considers data distortion factors such as uncertainty (noise in feature locations), occlusion (missing features), and clutter (spurious features). In addition, it considers structural similarity between model objects. The framework consists of two stages. In the first stage, we calculate a measure of the structural similarity between every pair of objects in the model set. This measure is a function of the relative transformation between the model objects. In the second stage, the model similarity information is used along with statistical models of the data distortion factors to determine bounds on the probability of correct recognition. The proposed framework is compared with relevant research efforts. Its validity is demonstrated using real synthetic aperture radar (SAR) data from the MSTAR public domain, which are obtained under a variety of depression angles and object configurations.
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
Arman, F., Aggarwal, J.: Model-based object recognition in dense-range images: A review. ACM Comput. Surveys 25 (1993) 5–43
Besl, P., Jain, R.: Three-dimensional object recognition. ACM Comput. Surveys 17 (1985) 75–145
Chin, R., Dyer, C.: Model-based recognition in robot vision. ACM Comput. Surveys 18 (1986) 67–108
Suetens, P., Fua, P., Hanson, A.: Computational strategies for object recognition. ACM Comput. Surveys 24 (1992) 5–61
Bhanu, B., III, G.J.: Recognizing target variants and articulations in synthetic aperture radar images. Optical Engineering 39 (2000) 712–723
Boshra, M., Zhang, H.: A constraint-satisfaction approach for 3-D object recognition by integrating 2-D and 3-D data. Comput. Vision Image Understand. 73 (1999) 200–214
Wells, W.: Statistical approaches to feature-based object recognition. Int. J. of Computer Vision 21 (1997) 63–98
Alter, T., Grimson, W.: Verifying model-based alignments in the presence of uncertainty. In: Proc. IEEE Conf. Comput. Vision and Patt. Recogn., San Juan, Puerto Rico (1997) 344–349
Boykov, Y., Huttenlocher, D.: A new Bayesian framework for object recognition. In: Proc. IEEE Conf. Comput. Vision and Patt. Recogn. Volume 2., Fort Collins, Colorado (1999) 517–523
III, G.J., Bhanu, B.: Recognition of articulated and occluded objects. IEEE Transactions on Pattern Anal. and Mach. Intell. 21 (1999) 603–613
Dhome, M., Richetin, M., Lapreste, J., Rives, G.: Determination of the attitude of 3-D objects from a single perspective view. IEEE Transactions on Pattern Anal. and Mach. Intell. 11 (1989) 1265–1278
Huttenlocher, D., Ullman, S.: Recognizing solid objects by alignment with an image. Int. J. of Computer Vision 5 (1990) 195–212
Dhome, M., Kasvand, T.: Polyhedra recognition by hypothesis accumulation. IEEE Trans. on Pattern Anal. and Mach. Intell. 9 (1987) 429–439
Stockman, G.: Object recognition and localization via pose clustering. Comput. Vision Graphics Image Process. 40 (1987) 361–387
Faugeras, O., Hebert, M.: The representation, recognition and locating of 3-D objects. Int. J. of Robotic Res. 5 (1986) 27–52
Grimson, W., Lozano-Perez, T.: Localizing overlapping parts by searching the interpretation tree. IEEE Trans. on Pattern Anal. and Mach. Intell. 9 (1987) 469–482
Grimson, W., Huttenlocher, D.: On the verification of hypothesized matches in model-based recognition. IEEE Transactions on Pattern Anal. and Mach. Intell. 13 (1991) 1201–1213
Sarachik, K.: The effect of Gaussian error in object recognition. IEEE Transactions on Pattern Anal. and Mach. Intell. 19 (1997) 289–301
Lindenbaum, M.: An integrated model for evaluating the amount of data required for reliable recognition. IEEE Transactions on Pattern Anal. and Mach. Intell. 19 (1997) 1251–1264
Irving, W., Washburn, R., Grimson, W.: Bounding performance of peak-based target detectors. In: Proc. SPIE Conference on Algorithms for Synthetic Aperture Radar Imagery IV. Volume 3070. (1997) 245–257
Lindenbaum, M.: Bounds on shape recognition performance. IEEE Transactions on Pattern Anal. and Mach. Intell. 17 (1995) 665–680
Grenander, U., Miller, M., Srivastava, A.: Hilbert-Schmidt lower bounds for estimators on matrix lie groups for ATR. IEEE Transactions on Pattern Anal. and Mach. Intell. 20 (1998) 790–802
Boshra, M., Bhanu, B.: Predicting performance of object recognition. IEEE Trans. on Pattern Anal. and Mach. Intell. 22 (2000) 956–969
Boshra, M., Bhanu, B.: Validation of SAR ATR performance prediction using learned distortion models. In: Proc. SPIE Conference on Algorithms for Synthetic Aperture Radar Imagery VII. Volume 4053., Orlando, Florida (2000) 558–566
Boshra, M., Bhanu, B.: Predicting an upper bound on SAR ATR performance. IEEE Trans. on Aerospace and Electronic Syst. 37 (2001) 876–888
Li, S.: Markov Random Field Modeling in Image Analysis. Springer-Verlag, New York (2001)
Ying, Z., Castanon, D.: Feature-based object recognition using statistical occlusion models with one-to-one correspondence. In: Proc. Int. Conf. on Comput. Vision. Volume 1., Vancouver, Canada (2001) 621–627
Ross, T., Worrell, S., Velten, V., Mossing, J., Bryant, M.: Standard SAR ATR evaluation experiments using the MSTAR public release data set. In: Proc. SPIE Conference on Algorithms for Synthetic Aperture Radar Imagery V. Volume 3370., Orlando, Florida (1998) 566–573
Pope, A., Lowe, D.: Probabilistic models of appearance for 3-D object recognition. Int. J. of Computer Vision 40 (2000) 149–167
Bhanu, B., Lin, Y., Jones, G., Peng, J.: Adaptive target recognition. Machine Vision and Applications 11 (2000) 289–299
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2005 Springer-Verlag London Limited
About this chapter
Cite this chapter
Boshra, M., Bhanu, B. (2005). A Theoretical Framework for Predicting Performance of Object Recognition. In: Bhanu, B., Pavlidis, I. (eds) Computer Vision Beyond the Visible Spectrum. Advances in Pattern Recognition. Springer, London. https://doi.org/10.1007/1-84628-065-6_1
Download citation
DOI: https://doi.org/10.1007/1-84628-065-6_1
Publisher Name: Springer, London
Print ISBN: 978-1-85233-604-2
Online ISBN: 978-1-84628-065-8
eBook Packages: Computer ScienceComputer Science (R0)