Object Recognition Using Log-Euclidean Multivariate Gaussian Descriptors

Visual features such as color, texture and shape are used extensively to represent image signatures in content based image retrieval systems (CBIR). The retrieval quality is adequate for some retrieval tasks but there is still a semantic gap between the low-level visual features (textures, shapes, colors) and automatically extracted high-level concepts that users normally search for. We proposed novel approach based on Log-Euclidean Multivariate Gaussian to represent the image visual features efficiently which enjoy the important applications of computer vision. This approach extracts image features using SIFT algorithm and these features again represented using multivariate Gaussian. The resulting descriptors called Log-Euclidean Multivariate Gaussian Descriptors, works well with low and high dimensional raw features. Extensive experiments were conducted to evaluate thoroughly this approach and the result showed that this approach is very competitive in object recognition and classification.

[1]  King-Sun Fu,et al.  Shape Discrimination Using Fourier Descriptors , 1977, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  D. Le Bihan,et al.  Diffusion tensor imaging: Concepts and applications , 2001, Journal of magnetic resonance imaging : JMRI.

[3]  B. Hall.  Lie Groups, Lie Algebras, and Representations , 2003 .

[4]  Yan Ke,et al.  PCA-SIFT: a more distinctive representation for local image descriptors , 2004, CVPR 2004.

[5]  David G. Lowe,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004, International Journal of Computer Vision.

[6]  Xavier Pennec,et al.  A Riemannian Framework for Tensor Computing , 2005, International Journal of Computer Vision.

[7]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[8]  Cordelia Schmid,et al.  Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[9]  Nicholas Ayache,et al.  Geometric Means in a Novel Vector Space Structure on Symmetric Positive-Definite Matrices , 2007, SIAM J. Matrix Anal. Appl..

[10]  Fang Liu,et al.  Shape of Gaussians as feature descriptors , 2009, CVPR.

[11]  Koen E. A. van de Sande,et al.  Evaluating Color Descriptors for Object and Scene Recognition , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  T. Gopal,et al.  Image retrieval using adapted Fourier Descriptors , 2010 .

[13]  Yoshua Bengio,et al.  Random Search for Hyper-Parameter Optimization , 2012, J. Mach. Learn. Res..

[14]  Thomas Mensink,et al.  Image Classification with the Fisher Vector: Theory and Practice , 2013, International Journal of Computer Vision.

[15]  Rita Cucchiara,et al.  Modeling local descriptors with multivariate gaussians for object and scene recognition , 2013, ACM Multimedia.

[16]  Takumi Kobayashi,et al.  Dirichlet-Based Histogram Feature Transform for Image Classification , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[17]  Gang Li,et al.  Stable Embedding of Grassmann Manifold via Gaussian Random Matrices , 2013, IEEE Transactions on Information Theory.

[18]  Lei Zhang,et al.  Local Log-Euclidean Multivariate Gaussian Descriptor and Its Application to Image Classification , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.