基于多条件随机场模型的图像3D空间布局理解

doi:10.3969/j.issn.0372-2112.2017.02.010

电子学报 ›› 2017, Vol. 45 ›› Issue (2): 328-336.DOI: 10.3969/j.issn.0372-2112.2017.02.010

基于多条件随机场模型的图像3D空间布局理解

刘威^1,2, 周婷², 袁淮¹, 赵宏¹

1. 东北大学研究院, 辽宁沈阳 110819;
2. 东软集团股份有限公司, 辽宁沈阳 110179

收稿日期:2015-06-12 修回日期:2015-12-09 出版日期:2017-02-25
- 作者简介:
- 刘威,男,1975年6月出生,辽宁沈阳人,博士,东北大学副教授,教授级高级工程师.主要研究领域为汽车辅助驾驶、智能交通、智能安防,研究方向为计算机视觉、图像处理、模式识别.E-mail:lwei@neusoft.com;周婷,女,1990年2月出生,四川乐山人,东软汽车电子研究员.主要研究领域为汽车辅助驾驶、自动驾驶,研究方向机器学习.E-mail:zhou_ting@neusoft.com
- 基金资助:
- 国家自然科学基金 (No.61273239）; 中央高校基本业务费 (No.N151802001）

3D Spatial Layout Understanding from Image Based on Multiple CRFs

LIU Wei^1,2, ZHOU Ting², YUAN Huai¹, ZHAO Hong¹

1. Research Academy, Northeastern University, Shenyang, Liaoning 110819, China;
2. Neusoft Corporation, Shenyang, Liaoning 110179, China

Received:2015-06-12 Revised:2015-12-09 Online:2017-02-25 Published:2017-02-25
- Supported by:
- National Natural Science Foundation of China (No.61273239); Fundamental Research Funds for the Central Universities (No.N151802001)

摘要/Abstract

摘要：

图像3D空间布局理解在自动驾驶系统以及目标识别中扮演着重要的角色.本文提出一种基于多条件随机场模型集成的图像3D空间布局理解算法.首先，基于多次图像分割生成多个不同尺度的超像素图像；然后，结合LBP表面纹理特征、LM滤波器组获得的方向纹理特征、颜色特征以及图像中超像素的位置和形状特征，建立各尺度的超像素图像中超像素的特征表达；最后，为各尺度的超像素图像分别构建相应的条件随机场模型，并应用D-S证据合成理论对多个条件随机场模型的推断结果进行集成，实现对图像3D空间布局的理解.在公共数据集GC和KITTI Layout上的实验结果表明，同已有算法相比，本文提出的算法提高了图像3D空间布局理解的准确率.

关键词: 3D空间布局, 多次图像分割, 超像素特征表达, 条件随机场模型, 证据合成

Abstract:

3D spatial layout understanding from images plays an important role in the autonomous driving and object recognition.This paper proposes a 3D spatial layout understanding algorithm based on multiple CRFs.Firstly,multiple different scales super-pixel image are generated based on the multiple segmentation.Then,the feature of super-pixel are constructed based on LBP surface texture feature,orientation texture feature from LM filters,color feature,and location and shape feature of super-pixels in the image.Finally,the CRF model on every scale super-pixel image is built,the Dempster-Shafer theory of evidence is used to integrate the inference result of multiple CRF models and the 3D spatial layout understanding from an image is realized.The experiments on the public database Geometric Context and KITTI Layout demonstrate that the algorithm proposed in this paper improves the average accuracy of 3D spatial layout understanding comparing to the existing state-of-art.

Key words: 3D spatial layout, multiple image segmentation, feature of super-pixel, conditional random field models, evidence combine

中图分类号:

TP391

刘威, 周婷, 袁淮, 等. 基于多条件随机场模型的图像3D空间布局理解[J]. 电子学报, 2017, 45(2): 328-336.

LIU Wei, ZHOU Ting, YUAN Huai, et al. 3D Spatial Layout Understanding from Image Based on Multiple CRFs[J]. Acta Electronica Sinica, 2017, 45(2): 328-336.

参考文献

[1] Alvarez J M,Gevers T,Lopez A M.3D scene priors for road detection[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR)[C].San Francisco,CA:IEEE Press,2010.57-64.
[2] Alvarez J M,Lopez AM,Gevers T.Combining priors,appearance,and context for road detection[J].IEEE Transactions on Intelligent Transportation Systems,2014,15(3):1168-1178.
[3] Hai W,Chao-chun Y,Ying-feng C.Smart road vehicle sensing system based on monocular vision[J].Journal for Light and Electron Optics,2015,126(4):386-390.
[4] Hoiem D,Efros A A,Hebert M.Geometric context from a single image[A].Proceedings of Tenth IEEE International Conference on Computer Vision (ICCV)[C].USA:IEEE,2005.654-661.
[5] Hoiem D,Efros A A,Hebert M.Recovering surface layout from an image[J].International Journal of Computer Vision,2007,75(1):151-172.
[6] Hoiem D,Efros A A,Hebert M.Closing the loop in scene interpretation[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition[C].Anchorage,AK:IEEE Press,2008.1-8.
[7] Divvala S K,Efros A A,Hebert M.Can similar scenes help surface layout estimation[A].Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops[C].USA:IEEE Press,2008.1-8.
[8] Divvala S K,Efros A A,Hebert M.Unsupervised Patch-Based Context From Millions of Images[R].Technology Report CMU-RI-TR-11-38,Robotics Institute:Carnegie Mellon University,2011.
[9] Gould S,Fulton R,Koller D.Decomposing a scene into geometric and semantically consistent regions[A].Proceedings of IEEE 12th International Conference on Computer Vision[C].Kyoto:IEEE Press,2009.1-8.
[10] Lazebnik S,Raginsky M.An empirical Bayes approach to contextual region classification[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition[C].Miami,FL:IEEE Press,2009.2380-2387.
[11] Lafferty J,McCallum A,Pereira F.Conditional random fields:Probabilistic models for segmentation and labeling sequence data[A].Proceedings of International Conference on Machine Learning[C].USA:Morgan Kaufmann,2001.282-289.
[12] Zhi-Hua Zhou.Ensemble Methods:Foundations and Algorithms[M].USA:CRC Press,2012.
[13] 姚旭,王晓丹,张玉玺.基于随机子空间和Adaboost的自适应集成方法[J].电子学报,2013,41(4):810-814. YAO Xu,WANG Xiao dan,ZHANG Yu xi.A self-adaption ensemble algorithm based on random subspace and Adaboost[J].Acta Electronica Sinica,2013,41(4):810-814.(in Chinese)
[14] 毕凯,王晓丹,姚旭.一种基于Bagging 和混淆矩阵的自适应选择性集成[J].电子学报,2014,42(4):711-716. BI Kai,WANG Xiaodan,YAO Xu.Adaptively selective ensemble algorithm based on bagging and confusion matrix[J].Acta Electronica Sinica,2014,42(4):711-716.(in Chinese)
[15] 朱鹏飞,等.基于随机化属性选择和邻域覆盖约简的集成学习[J].电子学报,2012,40(2):273-279. ZHU Peng-fei,et al.Ensemble learning based on randomized attribute selection and neighborhood covering reduction[J].Acta Electronica Sinica,2012,40(2):273-279.(in Chinese)
[16] Ping Zhong,Runsheng Wang.A multiple conditional random fields ensemble model for urban area detection in remote sensing optical images[J].IEEE Transactions on Geoscience and Remote Sensing,2007,12(45):3978-3988.
[17] Stricker M,Orengo M.Similarity of color images[J].Proceedings of SPIE Storage and Retrieval for Image and Video Databases,1995,2420:381-392.
[18] Thomas L,Jitendra M.Representing and recognizing the visual appearance of materials using three-dimensional textons[J].International Journal of Computer Vision,2001,43(1):29-44.
[19] Ojala T,Pietik? inen M.Multi-resolution gray scale and rotation invariant texture classification with local binary patterns[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(7):971-987.
[20] Kosecka J,Zhang W.Video compass[A].Proceedings of ECCV[C].Berlin:Springer-Verlag,2002.1-15.
[21] Guo G,Li S Z,Chan K L.Face recognition by support vector machines[A].Proceedings of International Conference on Automatic Face and Gesture Recognition[C].Grenoble:IEEE Press,2000.196-201.
[22] Liu Cheng-Lin.Classifier combination based on confidence transformation[J].Pattern Recgonition,2005,38(1):11-28.
[23] Bishop C M.Neural Networks for Pattern Recognition[M].USA:Oxford University Press,2004.
[24] Jordan M I,Ghahramani Z,Jaakkola T S,et al.An introduction to variational methods for graphical models[J].Machine Learning,1999,37(2):183-233.
[25] Shafer G.A Mathematical Theory of Evidence[M].Princeton,NJ:Princeton University Press,1976.
[26] José M álvarez,Yann LeCun,Theo Gevers,Antonio M López.Road scene segmentation from a single image[A].Proceedings of European Conference on Computer Vision (ECCV)[C].Florence:Springer-Verlag,2012.1-14.
[27] 秦锋,杨波,程泽凯.分类器性能评价标准研究[J].计算机技术与发展,2006,16(10):85-88. Qin Feng,Yang Bo,Cheng Ze-kai.Research on measure criteria in evaluating classification performance[J].Computer Technology and Development,2006,16(10):85-88.(in Chinese)
[28] 何周灿,王庆.一种用于图像匹配的快速有效的二分哈希搜索算法[J].西北工业大学学报,2010,28(4):609-615. He Zhoucan,Wang Qing.A fast and effective dichotomy-based hash (DBH) search algorithm for image matching[J].Journal of Northwestern Polytechnical Universty,2010,28(4):609-615.(in Chinese)
[29] 盛聚,谢千式.概率论与数理统计[M].北京:高等教育出版社,2008.

基于多条件随机场模型的图像3D空间布局理解

3D Spatial Layout Understanding from Image Based on Multiple CRFs

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	韩光洁, 赵腾飞, 刘立, 张帆, 徐政伟. 基于多元区域集划分的工业数据流概念漂移检测[J]. 电子学报, 2023, (): 1-11.
[2]	彭锦佳, 王辉兵. 基于异构卷积神经网络集成的无监督行人重识别方法[J]. 电子学报, 2023, (): 1-13.
[3]	余伶俐, 易倩, 金鸣岳, 周开军. 面向仿射目标识别的几何与仿生融合特征提取方法[J]. 电子学报, 2023, (): 1-12.
[4]	郭凯红, 崔明茜, 刘婷婷. 模糊知识测度下图像脉冲噪声去除方法[J]. 电子学报, 2023, (): 1-14.
[5]	李杨帅, 彭斐, 韩倩, 李小帅, 解光军. 一种针对QCA电路自动布局布线的混合策略研究[J]. 电子学报, 2023, 51(3): 666-674.
[6]	吕杭, 蒋明峰, 李杨, 张鞠成, 王志康. 基于混合时频域特征的卷积神经网络心律失常分类方法的研究[J]. 电子学报, 2023, 51(3): 701-711.
[7]	郑云飞, 王晓兵, 张雄伟, 曹铁勇, 孙蒙. 基于金字塔知识的自蒸馏HRNet目标分割方法[J]. 电子学报, 2023, 51(3): 746-756.
[8]	隗昊, 唐焕玲, 周爱, 张益嘉, 陈飞, 鲁明羽. 基于双路分段注意力神经张量网络的临床文本关系抽取[J]. 电子学报, 2023, 51(3): 658-665.
[9]	黄赟, 张帆, 郭威, 陈立, 羊光. 一种基于数据标准差的卷积神经网络量化方法[J]. 电子学报, 2023, 51(3): 639-647.
[10]	范兵兵, 何庭建, 张聪炫, 陈震, 黎明. 联合遮挡约束与残差补偿的特征金字塔光流计算方法[J]. 电子学报, 2023, 51(3): 648-657.
[11]	张晶, 王翌歆, 任永功. 统一全局空间表达的脑电信号跨被试情感识别[J]. 电子学报, 2023, (): 1-9.
[12]	申铉京, 李涵宇, 黄永平, 王玉. 基于自适应多尺度特征融合网络的车辆检测方法[J]. 电子学报, 2023, (): 1-9.
[13]	陈习坤, 杨俊美. 基于离散小波包变换与胶囊生成对抗网络的语音超分辨率算法[J]. 电子学报, 2023, (): 1-11.
[14]	张娜, 包梓群, 罗源, 吴彪, 涂小妹. 基于改进Cascade R-CNN算法在目标检测上的应用[J]. 电子学报, 2023, (): 1-11.
[15]	但志平, 方帅领, 孙航, 李晶, 万俊. 基于双判别器异构CycleGAN框架下多阶通道注意力校准的室外图像去雾[J]. 电子学报, 2023, (): 1-14.