计算机应用 ›› 2015, Vol. 35 ›› Issue (10): 2715-2720.DOI: 10.11772/j.issn.1001-9081.2015.10.2715

• 第十五届中国机器学习会议(CCML2015)论文 •    下一篇

MTRF:融合空间信息的主题模型

潘智勇1,2, 刘扬1, 刘国军1, 郭茂祖1, 李盼1   

  1. 1. 哈尔滨工业大学 计算机科学与技术学院, 哈尔滨 150001;
    2. 北华大学 信息技术与传媒学院, 吉林 吉林 132013
  • 收稿日期:2015-06-15 修回日期:2015-06-30 出版日期:2015-10-10 发布日期:2015-10-14
  • 通讯作者: 郭茂祖(1966-),男,山东夏津人,教授,博士生导师,博士,CCF会员,主要研究方向:机器学习、生物信息学,maozuguo@hit.edu.cn
  • 作者简介:潘智勇(1980-),男,吉林吉林人,实验师,博士研究生,主要研究方向:机器学习;刘扬(1976-),男,黑龙江哈尔滨人,副教授,博士,CCF会员,主要研究方向:机器学习、图像处理;刘国军(1979-),男,黑龙江哈尔滨人,讲师,博士,主要研究方向:机器学习;李盼(1988-),男,河北邢台人,硕士,主要研究方向:机器学习、统计学习。
  • 基金资助:
    国家自然科学基金资助项目(61171185,61271346);黑龙江省青年科学基金资助项目(QC2014C071)。

MTRF: a topic model with spatial information

PAN Zhiyong1,2, LIU Yang1, LIU Guojun1, GUO Maozu1, LI Pan1   

  1. 1. School of Computer Science and Technology, Harbin Institute of Technology, Harbin Heilongjiang 150001, China;
    2. College of Information Technology and Media, Beihua University, Jilin Jilin 132013, China
  • Received:2015-06-15 Revised:2015-06-30 Online:2015-10-10 Published:2015-10-14

摘要: 针对主题模型中词汇独立性和主题独立性假设忽略了视觉词汇间空间关系的问题,提出了一种融合了视觉词汇空间信息的主题模型,称为马尔可夫主题随机场(MTRF),并且提出了主题在图像处理中的表现形式为对象的组成部件。根据相邻视觉词汇以很大概率产生于同一主题的特点,该算法在产生主题的过程中,通过视觉词汇间是否产生于同一主题,来判断主题产生于马尔可夫随机场(MRF),还是产生于多项式分布。同时,从理论和实验两方面论证了主题并非对象的实例,而是以中层特征的形式表达对象的各个组成部件。与隐狄利克雷分配(LDA)相比,MTRF在Caltech101上的平均准确率提高了3.91%;在VOC2007数据集上的平均精度均值(mAP)提高了2.03%;此外,MTRF更准确地为视觉词汇分配了主题,能产生更有效表达对象的组成部件的中层特征。实验结果表明,MTRF有效地利用了空间信息,提高了模型的准确率。

关键词: 主题模型, 隐狄利克雷分配模型, 马尔可夫随机场, 空间关系, 中层特征, 图像分类

Abstract: To overcome the limitation of the assumptions of topic model-word independence and topic independence, a topic model which inosculated the spatial relationship of visual words was proposed, namely Markov Topic Random Field (MTRF). In addition, it was discussed that the "topic" of topic model represented the part of object in image processing. There is a high probability of the neighbor visual words generated from the same topic, and whether the visual words were generated from the same topic determined the topic was generated from Markov Random Field (MRF) or multinomial distribution of topic model. Meanwhile, both theoretical analysis and experimental results prove that "topic" of topic model appeared as mid-level feature to represent the parts of objects rather than the instances of objects. In experiments of image classification, the average accuracy of MTRF was 3.91% higher than that of Latent Dirichlet Allocation (LDA) on Caltech101 dataset, and the mean Average Precision (mAP) of MTRF was 2.03% higher than that of LDA on VOC2007 dataset. Furthermore, MTRF assigned topics to visual words more accurately and got the mid-level features which represented the parts of objects more effectively than LDA. The experimental results show that MTRF makes use of the spatial information effectively and improves the accuracy of the model.

Key words: topic model, Latent Dirichlet Allocation (LDA) model, Markov Random Field (MRF), spatial relationship, mid-level feature, image classification

中图分类号: