Abstract In the obstacle detection, obstacle/non-obstacle samples have characteristics of a large range of overlapping in the feature space and uneven distribution. The traditional training method for the classifier is not competent for dealing with such data. Thus, an improved method of maximizing area under the ROC (AUC) is proposed to train classifier. An alternative function is used as the objective function of optimizing AUC. Meanwhile, the particle swarm optimization is introduced to optimize the AUC objective function, and the particle swarm optimization algorithm is improved by using the Butterworth curves and particles with the low fitness value being mutated. The experimental results show that the proposed method effectively solves the local optimization caused by the gradient descent method. Moreover, the detection performance of the proposed method is improved compared with other existing algorithms, and the algorithm is reliable and efficient.
[1] Adler W, Lausen B. Bootstrap Estimated True and False Positive Rates and ROC Curve. Computational Statistics and Data Analysis, 2009, 53(3): 718-729 [2] Alsing S G, Bauer K W, Miller J O. A Multinomial Selection Procedure for Evaluating Pattern Recognition Algorithms. Pattern Recognition, 2002, 35(11): 2397-2412 [3] Flach P A. The Many Faces of ROC Analysis in Machine Learning // Proc of the 21st International Conference on Machine Learning. Alberta, Canada, 2004: 485-487 [4] Cortes C, Mohri M. AUC Optimization vs. Error Rate Minimization // Proc of the 18th Annual Conference on Neural Information Processing Systems. Vancouver, Canada, 2004: 16-24 [5] Ling C X, Huang Jin, Zhang H. AUC: A Better Measure Than Accuracy in Comparing Learning Algorithms // Proc of the International Joint Conferences on Artificial Intelligence. Acapulco, Mexico, 2003: 329-341 [6] Provost F J, Fawcett T, Kohavi R. The Case against Accuracy Estimation for Comparing Induction Algorithms // Proc of the 15th International Conference on Machine Learning. Wisconsin, USA, 1998: 445-453 [7] Tax D M J, Veenman C J. Tuning the Hyperparameter of an AUC-Optimized Classifier // Proc of the 17th Belgium-Netherlands Conference on Artificial Intelligence. Brussels, Belgium, 2005: 224-231 [8] Joachims T. A Support Vector Method for Multivariate Performance Measures // Proc of the 22nd International Conference on Machine Learning. Bonn, Germany, 2005: 377-384
[9] Tax D J M, Duin R P W, Arzhaeva Y. Linear Model Combining by Optimizing the Area under the ROC Curve // Proc of the 18th IEEE International Conference on Pattern Recognition. Hongkong, China, 2006: 119-122 [10] Marrocco C, Molinara M, Tortorella F. Exploiting AUC for Optimal Linear Combination of Dichotomizers. Pattern Recognition Letters, 2006, 27(8): 900-907 [11] Toh K A, Kim J, Lee S. Maximizing Area under ROC Curve for Biometric Scores Fusion. Pattern Recognition, 2008, 41(11): 3373-3392 [12] Marroccoa C, Duinb R P W, Tortorellaa F. Maximizing the Area under the ROC Curve by Pairwise Feature Combination. Pattern Recognition, 2008, 41(6): 1961-1974 [13] Yan Lian, Dodier R, Mozer M C, et al. Optimizing Classifier Performance via the Wilcoxon-Mann-Whitney Statistics // Proc of the 20th International Conference on Machine Learning. Washington, USA, 2003: 848-855 [14] Calders T, Jaroszewicz S. Efficient AUC Optimization for Classification // Proc of the 11th European Conference on Principles and Practice of Knowledge Discovery in Databases. Warsaw, Poland, 2007: 42-53 [15] Herschtal A, Raskutti B. Optimising Area under the ROC Curve Using Gradient Descent // Proc of the 21st International Conference on Machine Learning. Alberta, Canada, 2004: 385-392 [16] Dima C, Hebert M. Active Learning for Outdoor Obstacle Detection // Proc of the Robotics: Science and Systems Conference. Cambridge, USA, 2005: 9-16 [17] Manduchi R, Castano A, Talukder A, et al. Obstacle Detection and Terrain Classification for Autonomous Off-Road Navigation. Autonomous Robots, 2005, 18(1): 81-102 [18] Fawcett T. An Introduction to ROC Analysis. Pattern Recognition Letters, 2006, 27(8): 861-874 [19] Eberhart R C, Shi Yuhui. Particle Swarm Optimization: Developments, Applications and Resources // Proc of the Congress on Evolutionary Computation. Seoul, South Korea, 2001, Ⅰ: 81-86 [20] Chen Yunjie, Zhang Jianwei, Wei Zhihui, et al. Brain MRI Segmentation Using the Active Contours Based on Gaussian Mixture Models. Journal of Computer Research and Development, 2007, 44(9): 1595-1603 (in Chinese) (陈允杰,张建伟,韦志辉,等.基于高斯混合模型的活动轮廓模型脑MRI分割.计算机研究与发展, 2007, 44(9): 1595-1603) [21] Permuter H, Jermyn J. A Study of Gaussian Mixture Models of Color and Texture Features for Image Classification and Segmentation. Pattern Recognition, 2006, 39(4): 695-706 [22] Castano R, Manduchi R, Fox J. Classification Experiments on Real-World Texture // Proc of the 3rd Workshop on Empirical Evaluation Methods in Computer Vision. Hawaii, USA, 2001: 3-20
2010, Vol. 23 
