IMPROVING OBJECT DETECTION PERFORMANCE WITH GENETIC PROGRAMMING
Abstract
This paper describes three developments to improve object detection performance using genetic programming. The first investigates three feature sets, the second investigates a new fitness function, and the third introduces a two phase learning method using genetic programming. This approach is examined on three object detection problems of increasing difficulty and compared with a neural network approach. The two phase GP approach with the new fitness function and the local concentric circular region features achieved the best results. The results suggest that the concentric circular pixel statistics are more effective than the square features for these object detection problems. The fitness function with program size is more effective and more efficient than without for these object detection problems and the evolved genetic programs using this fitness function are much shorter and easier to interpret. The two phase GP approach is more effective and more efficient than the single stage GP approach, and also more effective than the neural network approach on these problems using the same set of features.
References
-
John R. Koza , Genetic programming : on the programming of computers by means of natural selection ( MIT Press , Cambridge, Mass., London, England , 1992 ) . Google Scholar -
Wolfgang Banzhaf , Genetic Programming: An Introduction on the Automatic Evolution of computer programs and its Applications ( Morgan Kaufmann Publishers, Dpunkt-verlag , San Francisco, Calif., Heidelburg , 1998 ) . Crossref, Google Scholar Walter Alden Tackett , Genetic programming for feature discovery and image discrimination. In Stephanie Forrest, editor, Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 (Morgan Kaufmann, 1993) pp. 303–309. Google ScholarKarl Benson , Evolving finite state machines with embedded genetic programming for automatic target detection within SAR imagery, Proceedings of the 2000 Congress on Evolutionary Computation CEC00 (IEEE Press, 2000) pp. 1543–1549. Google Scholar- Real-World Applications of Evolutionary Computing,
LNCS 1803, eds.Stefano Cagnoni (Springer-Verlag, Edinburgh, 2000) pp. 12–21. Crossref, Google Scholar , Daniel Howard , Simon C. Roberts and Conor Ryan , The boru data crawler for object detection tasks in machine vision, Applications of Evolutionary Computing, Proceedings of EvoWorkshops2002: EvoCOP, EvoIASP, EvoSTim2279,LNCS , eds.Stefano Cagnoni (Springer-Verlag) pp. 220–230. Google Scholar-
F. Lindblad , P. Nordin and K. Wolff , Evolving 3d model interpretation of images using graphics hardware , Proceedings of the 2002 IEEE Congress on Evolutionary Computation, CEC2002 ( 2002 ) . Google Scholar Jamie R. Sherrah , Robert E. Bogner and Abdesselam Bouzerdoum , The evolutionary pre-processor: Automatic feature extraction for supervised classification using genetic programming, Genetic Programming 1997: Proceedings of the Second Annual Conference, eds.John R. Koza (Morgan Kaufmann) pp. 304–312. Google ScholarMengjie Zhang and Victor Ciesielski , Genetic programming for multiple class object detection, Proceedings of the 12th Australian Joint Conference on Artificial Intelligence (AI'99)1747,Lecture Notes in Artificial Intelligence (LNAI) , ed.Norman Foo (Springer-Verlag Berlin Heidelberg, 1999) pp. 180–192. Google Scholar- Applications of Evolutionary Computing,
Lecture Notes in Computer Science, LNCS 2611, ed.Stefano Cagnoni (Springer-Verlag, 2003) pp. 455–466. Crossref, Google Scholar , - EURASIP Journal on Signal Processing, Special Issue on Genetic and Evolutionary Computation for Signal Processing and Image Analysis 2003(8), 841 (2003). Web of Science, Google Scholar
- Andy Song, Texture Classification: A Genetic Programming Approach. PhD thesis, Department of Computer Science, RMIT University, Melbourne, Australia, 2003 . Google Scholar
- IEEE Transactions on Neural Networks 6(1), 252 (1995), DOI: 10.1109/72.363430. Crossref, Web of Science, Google Scholar
- IEEE Transactions on neural networks 1(1), 28 (1990), DOI: 10.1109/72.80203. Crossref, Google Scholar
- Walter Alden Tackett, Recombination, Selection, and the Genetic Construction of Computer Programs. PhD thesis, Faculty of the Graduate School, University of Southern California, Canoga Park, California, USA, April 1994 . Google Scholar
- Advances in Genetic Programming, ed.
Kenneth E. Kinnear (MIT Press, 1994) pp. 477–494. Google Scholar , -
John R. Koza , Genetic Programming II: Automatic Discovery of Reusable Programs ( MIT Press , Cambridge, Mass., London, England , 1994 ) . Google Scholar Astro Teller and Manuela Veloso , A controlled experiment: Evolution for learning difficult image classification, Proceedings of the 7th Portuguese Conference on Artificial Intelligence990,LNAI , eds.Carlos Pinto-Ferreira and Nuno J. Mamede (Springer Verlag) pp. 165–176. Google ScholarStephen A. Stanhope and Jason M. Daida , Genetic programming for automatic target classification and recognition in synthetic aperture radar imagery, Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming1447,LNCS , eds.V. William Porto (Springer-Verlag) pp. 735–744. Google ScholarJay F. Winkeler and B. S. Manjunath , Genetic programming for object detection, Genetic Programming 1997: Proceedings of the Second Annual Conference, eds.John R. Koza (Morgan Kaufmann) pp. 330–335. Google Scholar- Astro Teller and Manuela Veloso, PADO: Learning tree structured algorithms for orchestration into an object recognition system. Technical Report CMU-CS-95-101, Department of Computer Science, Carnegie Mellon University, Pittsburgh, PA, USA, 1995 . Google Scholar
- Evolutionary Computation ,
Lecture Note in Computer Science 993 , ed.T. C. Fogarty ( Springer-Verlag , 1995 ) . Google Scholar , - Applied Soft Computing 4(2), 175 (2004), DOI: 10.1016/j.asoc.2004.01.004. Crossref, Web of Science, Google Scholar
- IEEE Transactions on Systems, Man and Cybernetics, Part B 35(3), 538 (2005), DOI: 10.1109/TSMCB.2005.846656. Crossref, Web of Science, Google Scholar
Mark E. Roberts and Ela Claridge , A multistage approach to cooperatively coevolving feature construction and object detection, Applications of Evolutionary Computing, EvoWorkshops2005: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoMUSART, EvoSTOC3449,LNCS , eds.Franz Rothlauf (Springer Verlag) pp. 396–406. Google ScholarMark E. Roberts and Ela Claridge , Cooperative coevolution of image feature construction and object detection, Parallel Problem Solving from Nature – PPSN VIII3242,LNCS , eds.Xin Yao (Springer-Verlag) pp. 902–911. Google ScholarMengjie Zhang and Malcolm Lett , Localisation fitness in GP for object detection, Applications of Evolutionary Computing, EvoWorkshops2006: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC3907,LNCS , eds.Franz Rothlauf (Springer Verlag) pp. 472–483. Google Scholar- Advances in Engineering Software 30, 303 (1999), DOI: 10.1016/S0965-9978(98)00093-3. Crossref, Web of Science, Google Scholar
Satoru Isaka , An empirical study of facial image feature extraction by genetic programming, the Genetic Programming 1997 Conference, ed.John R. Koza (Stanford Bookstore, Stanford University) pp. 93–99. Google ScholarMark E. Roberts , The effectiveness of cost based subtree caching mechanisms in typed genetic programming for image segmentation, Applications of Evolutionary Computing, EvoWorkshops2003: EvoBIO, EvoCOP, EvoIASP, EvoMUSART, EvoROB, EvoSTIM2611,LNCS , eds.Günther R. Raidl (University of Essex) pp. 444–454. Google ScholarMark E. Roberts and Ela Claridge , An artificially evolved vision system for segmenting skin lesion images, Proceedings of the 6th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI)2878,LNCS , eds.Randy E. Ellis and Terry M. Peters (Springer-Verlag, 2003) pp. 655–662. Google ScholarBradley J. Lucier , Sudhakar Mamillapalli and Jens Palsberg , Program optimisation for faster genetic programming, Genetic Programming – GP'98 pp. 202–207. Google ScholarJohn R. Koza , Simultaneous discovery of reusable detectors and subroutines using genetic programming, Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93, ed.Stephanie Forrest (Morgan Kauffman, 1993) pp. 295–302. Google ScholarDaniel Howard , Simon C. Roberts and Conor Ryan , Evolution of an object detection ant for image analysis, 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers, ed.Erik D. Goodman pp. 168–175. Google Scholar- Genetic Programming and Evolvable Machines 7(1), 81 (2007), DOI: 10.1007/s10710-006-7012-3. Crossref, Google Scholar
Riccardo Poli , Genetic programming for image analysis, Genetic Programming 1996: Proceedings of the First A nnual Conference, eds.John R. Koza (MIT Press, 1996) pp. 363–368. Google ScholarPeter Nordin and Wolfgang Banzhaf , Programmatic compression of images and sound, Genetic Programming 1996: Proceedings of the First Annual Conference, eds.John R. Koza (MIT Press, 1996) pp. 345–350. Google Scholar-
Olivier Faugeras , Three-Dimensional Computer Vision – A Geometric Viewpoint ( The MIT Press , 1993 ) . Google Scholar - Computer Vision and Image Understanding 63(3), 399 (1996). Crossref, Web of Science, Google Scholar
- Neural Networks 8(7), 1117 (1995), DOI: 10.1016/0893-6080(95)00047-X. Crossref, Web of Science, Google Scholar
- Neural Networks 8(7), 1153 (1995), DOI: 10.1016/0893-6080(95)00050-X. Crossref, Web of Science, Google Scholar
-
F. Samaria and A. Harter , Parameterisation of a stochastic model for human face identification , 2nd IEEE Workshop on Applications of Computer Vision . Google Scholar - Urvesh Bhowan, A. domain independent approach to multi-class object detection using genetic programming. BSc Honours research project, School of Mathematical and Computing Sciences, Victoria University of Wellington, 2003 . Google Scholar
- Bunna Ny., Multi-class object classification and detection using neural networks. BSc Honours research project, School of Mathematical and Computing Sciences, Victoria University of Wellington, 2003 . Google Scholar
Remember to check out the Most Cited Articles! |
---|
Check out Notable Titles in Artificial Intelligence. |