Image Semantic Distance Metric Learning Approach for Large-scale Automatic Image Annotation

Learning an effective semantic distance measure is very important for the practical application of image analysis and pattern recognition. Automatic image annotation (AIA) is a task of assigning one or more semantic concepts to a given image and a promising way to achieve more effective image retrieval and analysis. Due to the semantic gap between low-level visual features and high-level image semantic, the performances of some image distance metric learning (IDML) algorithms only using low-level visual features is not satisfactory. Since there is the diversity and complexity of large-scale image dataset, only using visual similarity to learn image distance is not enough. To solve this problem, in this paper, the semantic labels of the training image set participate into the image distance measure learning. The experimental results confirm that the proposed image semantic distance metric learning (ISDML) can improve the efficiency of large-scale AIA approach and achieve better annotation performance than the other state-of-the art AIA approaches.

[1]  Natsuda Kaothanthong,et al.  A feature-word-topic model for image annotation and retrieval , 2013, TWEB.

[2]  Tat-Seng Chua,et al.  NUS-WIDE: a real-world web image database from National University of Singapore , 2009, CIVR '09.

[3]  Changsheng Xu,et al.  Weakly Supervised Graph Propagation Towards Collective Image Parsing , 2012, IEEE Transactions on Multimedia.

[4]  Gang Chen,et al.  Semi-supervised Multi-label Learning by Solving a Sylvester Equation , 2008, SDM.

[5]  R. Manmatha,et al.  Multiple Bernoulli relevance models for image and video annotation , 2004, Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004..

[6]  Yannick Berthoumieu,et al.  Gaussian Copula Multivariate Modeling for Texture Image Retrieval Using Wavelet Transforms , 2014, IEEE Transactions on Image Processing.

[7]  Bernhard Schölkopf,et al.  Learning with Local and Global Consistency , 2003, NIPS.

[8]  Vladimir Pavlovic,et al.  A New Baseline for Image Annotation , 2008, ECCV.

[9]  C. V. Jawahar,et al.  Image Annotation Using Metric Learning in Semantic Neighbourhoods , 2012, ECCV.

[10]  Cong Jin,et al.  Automatic image annotation using feature selection based on improving quantum particle swarm optimization , 2015, Signal Process..

[11]  Yueting Zhuang,et al.  Apply semantic template to support content-based image retrieval , 1999, Electronic Imaging.

[12]  Changhu Wang,et al.  Learning to reduce the semantic gap in web image retrieval and annotation , 2008, SIGIR '08.

[13]  Cong Jin,et al.  Image Semantic Annotation Approach Based on the Feature Matching , 2014 .

[14]  Cordelia Schmid,et al.  TagProp: Discriminative metric learning in nearest neighbor models for image auto-annotation , 2009, 2009 IEEE 12th International Conference on Computer Vision.

[15]  中山 英樹.  Linear distance metric learning for large-scale generic image recognition , 2011 .

[16]  Jitendra Malik,et al.  Normalized cuts and image segmentation , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[17]  Michael Grubinger,et al.  Analysis and evaluation of visual information systems performance , 2007 .

[18]  Supavadee Aramvith,et al.  Two-Probabilistic Latent Semantic Model for Image Annotation and Retrieval , 2010, ACCV Workshops.

[19]  Geoffrey E. Hinton,et al.  Neighbourhood Components Analysis , 2004, NIPS.

[20]  Laura A. Dabbish,et al.  Labeling images with a computer game , 2004, AAAI Spring Symposium: Knowledge Collection from Volunteer Contributors.

[21]  Sally A. Goldman,et al.  MISSL: multiple-instance semi-supervised learning , 2006, ICML.

[22]  Cong Jin,et al.  A Hybrid Model Based on Mutual Information and Support Vector Machine for Automatic Image Annotation , 2015, CSOC.