Salient Building Region Detection for Scene Classification

Abstract

Abstract: In order to achieve outdoor scene classification in urban areas quickly, an effective method for detecting salient building region based on visual attention selection mechanism was proposed. Firstly, eightdirectional Gabor filter is used to obtain horizontal-vertical enhanced image. Secondly, salient building candidate regions are extracted through mapping from salient map which is constructed by phase Fourier tranform to the enhanced image. Finally, an orientation histogram discrimination algorithm is used to eliminate interference target, and salient building region detection is completed by CV model segmentation. The method extracts the building regional information which can represent scene based on PFT and orientation features, it effectively improves the real-time of algorithm and the capability of scene analysis, and implements the accurate detection of region in pixel level. The results of experiments with the university of Southern California outdoor scene database show that the method performs well in the detection result, computational speed and robustness.

Key words: saliency map, region detection, CV model, Gabor filter

CLC Number:

TP391.41

CHEN Shuo, YU Xiao-Sheng, WU Cheng-Dong, CHEN Dong-Yue. Salient Building Region Detection for Scene Classification[J]. Journal of Shanghai Jiaotong University, 2011, 45(08): 1130-1135.

References

［1］李振华, 敬忠良, 孙韶媛, 等. 基于目标检测的红外和可见光动态图像融合［J］. 上海交通大学学报, 2005, 39(8): 13041307.

LI Zhenhua, JING Zhongliang, SUN Shaoyuan, et al. Infrared and visible dynamic image fusion based on moving target detection［J］. Journal of Shanghai Jiaotong University, 2005, 39(8): 13041307.

［2］Song Dongjin, Tao Dacheng. Biologically inspired feature manifold for scene classification［J］. IEEE Transactions on Image Processing, 2010, 19(1): 174184.

［3］Lowe D G. Distinctive image features from scaleinvariant keypoints［J］. International Journal of Computer Vision, 2004, 60(2): 91110.

［4］Mikolajczyk K, Schmid C. Scale and affine invariant interest point detectors［J］. International Journal of Computer Vision, 2004, 60(1): 6386.

［5］Harris C, Stephens M J. A combined corner and edge detector［C］//Proceedings of the 4th Alvey Vision Conference. Manchester, U K: ［s.n.］, 1988: 147151.

［6］Kumar M P, Koller D. Efficiently selecting regions for scene understanding［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2010: 32173224.

［7］Walther D, Koch C. Modeling attention to salient protoobjects［J］. Neural Networks, 2006, 19(5): 13951407.

［8］Itti L, Koch C, Niebur E. A model of saliencybased visual attention for rapid scene analysis［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(11): 12541259.

［9］Neil D B, John K T. Saliency, attention, and visual search: An information theoretic approach［J］. Journal of Vision, 2009, 9(3): 124.

［10］Hou Xiaodi, Zhang Liqing. Saliency detection: A spectral residual approach［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2007: 24382445.

［11］Guo Chenlei, Ma Qi, Zhang Liming. Spatiotemporal saliency detection using phase spectrumof quaternion fourier transform［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2008: 18.

［12］Chen Dongyue, Han Peng, Wu Chengdong. Frequency spectrum modification: A new model for visual saliency detection［C］//Proceedings of the 7th International Symposium on Neural Networks. Berlin, Germany: Springer Verlag, 2010: 9096.

［13］Treisman A M, Gelade G.. A featureintegration theory of attention ［J］. Cognitive Psychology, 1980, 12(1): 97136.

［14］曲延云, 郑南宁, 李翠华, 等. 基于支持向量机的显著性建筑物检测［J］. 计算机研究与发展, 2007, 44(1): 141147.

QU Yanyun, ZHANG Nanning, LI Cuihua, et al. Salient building detection based on SVM［J］. Journal of Computer Research and Development, 2007, 44(1): 141147.