|本期目录/Table of Contents|

[1]邓燕子,卢朝阳,李静,等.采用多层图模型推理的道路场景分割算法[J].西安交通大学学报,2017,51(12):62-67.[doi:10.7652/xjtuxb201712010]
 DENG Yanzi,LU Zhaoyang,LI Jing,et al.A Segmentation Algorithm for Road Scenes Using Hierarchical GraphBased Inference[J].Journal of Xi'an Jiaotong University,2017,51(12):62-67.[doi:10.7652/xjtuxb201712010]
点击复制

采用多层图模型推理的道路场景分割算法(PDF)

《西安交通大学学报》[ISSN:0253-987X/CN:61-1069/T]

卷:
51
期数:
2017年第12期
页码:
62-67
栏目:
出版日期:
2017-12-10

文章信息/Info

Title:
A Segmentation Algorithm for Road Scenes Using
Hierarchical GraphBased Inference
作者:
邓燕子卢朝阳李静刘阳
西安电子科技大学综合业务网理论及关键技术国家重点实验室,710071,西安
Author(s):
DENG YanziLU ZhaoyangLI JingLIU Yang
State Key Laboratory of Integrated Service Network, Xidian University, Xi’an 710071, China
关键词:
道路场景分割多类别图像标记随机森林马尔科夫随机场
Keywords:
road scene segmentation multiclass image labeling random forest markov random field
分类号:
TP391.4
DOI:
10.7652/xjtuxb201712010
摘要:
针对传统图模型分割算法提取的物体边缘不够精细、难以适应复杂道路场景布局的问题,提出了一种基于多层图模型推理的道路场景分割(HGI)算法。该算法先将图像过分割为同质的超像素块,再采用随机森林模型训练超像素块的多类别回归器和相邻超像素的一致性回归器;然后用2种回归值计算马尔科夫随机场(MRF)模型的能量项,通过推理得到初始分割;最后为了解决超像素块包含多类别带来的分类混淆,在初始分割基础上构建像素级的全连接条件随机场模型,进行优化得到精细的分割结果。实验结果表明,采用HGI算法对人工标注数据库和真实拍摄的场景图像处理能够得到精细的分割边缘,能够解决超像素推理中的类别混淆问题,与传统的MRF图模型分割方法相比,在总体精度和平均召回率2个指标上分别提高了2%和3%。
Abstract:
A novel segmentation algorithm for road scenes based on hierarchical graphbased inference (HGI) is proposed to solve the problem that object boundaries extracted by existing graphbased segmentation algorithms are not fine enough and it is hard to adapt to complex road scene layouts. The algorithm first oversegments an image into small homogeneous regions called superpixels, and then the random forest model is used to train a multiclass regressor and a consistency regressor of superpixels. Regression results are then used to calculate the energy terms in a Markov random field (MRF) energy function. An initial segmentation of the image is obtained by using the superpixel MRF inference. A pixellevel labeling based on fully connected conditional random fields is constructed to avoid the label confusion caused by the superpixels and to get a fine segmentation finally. Experimental results show that the proposed algorithm solves the label confusion in the superpixel inference and gets fine segmentation boundaries for both the images in manually labeled datasets and the real road scenes. A comparison with the traditional MRF graphbased inference methods shows that the HGI algorithm provides improvements of 2% and 3% on the overall precision and perclass average metrics, respectively.

参考文献/References:

[1]HOIEM D, EFROS A, HEBERT M. Recovering surface layout from an image [J]. International Journal of Computer Vision, 2007, 75(1): 151172.
[2]COSTEA A D, NEDEVSCHI S. Fast traffic scene segmentation using multirange features from multiresolution filtered and spatial context channels [C]∥IEEE Intelligent Vehicles Symposium Proceedings. Piscataway, NJ, USA: IEEE, 2016: 328334.
[3]GOULD S. Darwin: a framework for machine learning and computer vision research and development [J]. Journal of Machine Learning Research, 2012, 13(18): 35333537.
[4]COSTEA A D, NEDEVSCHI S. Semantic channels for fast pedestrian detection [C]∥Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2016: 23602368.
[5]邓燕子, 卢朝阳, 李静. 交通场景的多视觉特征图像分割方法 [J]. 西安电子科技大学学报, 2015, 42(6): 1116.
DENG Yanzi, LU Zhaoyang, LI Jing. Segmentation of the image with multivisual features for a traffic scene [J]. Journal of Xidian University, 2015, 42(6): 1116.
[6]CHACRA D A, ZELEK J. Road segmentation in street view images using texture information [C]∥Proceedings of 2016 IEEE Conference on Computer and Robot Vision. Piscataway, NJ, USA: IEEE, 2016: 424431.
[7]FARABET C, COUPRIE C, NAJMAN L, et al. Learning hierarchical features for scene labeling [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(8): 19151929.
[8]王海, 蔡英凤, 贾允毅, 等. 基于深度卷积神经网络的场景自适应道路分割算法 [J]. 电子与信息学报, 2017, 39(2): 263269.
WANG Hai, CAI Yingfeng, JIA Yunyi, et al. Scene adaptive road segmentation algorithm based on deep convolutional neural network [J]. Journal of Electronics and Information Technology, 2017, 39(2): 263269.
[9]余淼, 胡占义. 高阶马尔科夫随机场及其在场景理解中的应用 [J]. 自动化学报, 2015, 41(7): 12131234.
YU Miao, HU Zhanyi. Higherorder Markov random fields and their applications in scene understanding [J]. Acta Automatica Sinica, 2015, 41(7): 12131234.
[10]HORNE L, ALVAREZ J M, SALZMANN M, et al. Efficient scene parsing by sampling unary potentials in a fullyconnected CRF [C]∥IEEE Intelligent Vehicles Symposium Proceedings. Piscataway, NJ, USA: IEEE, 2015: 820825.
[11]KRAHENBUHL P, KOLTUN V. Efficient inference in fully connected CRFs with Gaussian edge potentials [C]∥Proceedings of 25th Annual Conference on Neural Information Processing Systems 2011. Red Hook, NY, USA: Curran Associates Inc., 2011: 109117.
[12]FELZENSZWALB P F, HUTTENLOCHER D P. Efficient graphbased image segmentation [J]. International Journal of Computer Vision, 2004, 59(2): 167181.
[13]ARBELAEZ P, MAIRE M, FOWLKES C, et al. Contour detection and hierarchical image segmentation [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(5): 898916.
[14]XIAO L, DAI B, LIU D, et al. Monocular road detection using structured random forest [J]. International Journal of Advanced Robotic Systems, 2016, 13(3): 101114.
[15]徐胜军, 韩九强, 赵亮, 等. 用于图像分割的局部区域能量最小化算法 [J]. 西安交通大学学报, 2011, 45(8): 712.
XU Shengjun, HAN Jiuqiang, ZHAO Liang, et al. Algorithm of minimizing local region energy for image segmentation [J]. Journal of Xi’an Jiaotong University, 2011, 45(8): 712.
[16]GRIDCHYN I, KOLMOGOROV V. Potts model, parametric maxflow and Ksubmodular functions [C]∥Proceedings of the IEEE International Conference on Computer Vision. Piscataway, NJ, USA: IEEE, 2013: 23202327.
[17]ADAMS A, BAEK J, DAVIS M A. Fast highdimensional filtering using the permutohedral lattice [J]. Computer Graphics Forum, 2010, 29(2): 753762.
[18]ALVAREZ J M, GEVERS T, LECUN Y, et al. Road scene segmentation from a single image [C]∥Proceedings of the 12th European Conference on Computer Vision. Berlin, Germany: Springer Verlag, 2012: 376389.
[19]LADICKY L, RUSSELL C, KOHLI P, et al. Associative hierarchical random fields [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(6): 10561077.

备注/Memo

备注/Memo:
国家自然科学基金资助项目(61502364)
更新日期/Last Update: