One paper is online on Neurocomputing. Con~ to Hongli！

Hessian Regularization by Patch Alignment Framework

Abstract

In recent years, semi-supervised learning has played a key part in large-scale image management, where usually only a few images are labeled. To address this problem, many representative works have been reported, including transductive SVM, universum SVM, co-training and graph-based methods. The prominent method is the patch alignment framework, which unifies the traditional spectral analysis methods. In this paper, we propose Hessian regression based on the patch alignment framework. In particular, we construct a Hessian using the patch alignment framework and apply it to regression problems. To the best of our knowledge, there is no report on Hessian construction from the patch alignment viewpoint. Compared with the traditional Laplacian regularization, Hessian can better match the data and then leverage the performance. To validate the effectiveness of the proposed method, we conduct human face recognition experiments on a celebrity face dataset. The experimental results demonstrate the superiority of the proposed solution in human face classification.

Keywords

Semi-supervised learning; Hessian; Patch alignment; Least Squares

doi:10.1016/j.neucom.2015.07.152

TIE paper entitled “p-Laplacian Regularized Sparse Coding for Human Activity Recognition” is online.

p-Laplacian Regularized Sparse Coding for Human Activity Recognition

Authors

W. Liu 

Weifeng Liu is with the College of Information and Control Engineering, China University of Petroleum (East China), Qingdao 266580, China (email: liuwf@ upc.edu.cn). 

Z. J. Zha ; Y. Wang ; K. Lu ; D. Tao

Abstract

Human activity analysis in videos has increasingly attracted attention in computer vision research with the massive number of videos now accessible online. Although many recognition algorithms have been reported recently, activity representation is challenging. Recently, manifold regularized sparse coding has obtained promising performance in action recognition, because it simultaneously learns the sparse representation and preserves the manifold structure. In this paper, we propose a generalized version of Laplacian regularized sparse coding for human activity recognition called p-Laplacian regularized sparse coding. The proposed method exploits p-Laplacian regularization to preserve the local geometry. The p-Laplacian is a nonlinear generalization of standard graph Laplacian and has tighter isoperimetric inequality. As a result, p-Laplacian regularized sparse coding provides superior theoretical evidence than standard Laplacian regularized sparse coding with a proper p. We also provide a fast iterative Shrinkage-Thresholding algorithm (FISTA) for the optimization of p-Laplacian regularized sparse coding. Lastly, we input the sparse codes learned by the p- Laplacian regularized sparse coding algorithm into support vector machines and conduct extensive experiments on the unstructured social activity attribute (USAA) dataset and human motion database (HMDB51) for human activity recognition. The experimental results demonstrate that the proposed p-Laplacian regularized sparse coding algorithm outperforms the manifold regularized sparse coding algorithms including the standard Laplacian regularized sparse coding algorithm with a proper p.

Keywords

human activity recognition

manifold

p-Laplacian

sparse coding

DOI:

10.1109/TIE.2016.2552147

CFP: IEEE SMC2016 special session on Cognitive Computing

http://www.smc2016.org/

Special Session Call for Papers

SMC2016 Special Session on Matrix Analysis and Feature Learning for Multimedia Understanding

Introduction/Call for Papers

Matrix analysis and feature learning have important applications in natural sciences, medicine, economics, engineering as well as in industry. The research of matrix analysis and feature learning plays important roles in dimension reduction, spectral analysis, manifold learning, kernel machines, sparse coding, pattern recognition etc. In recent years, the matrix analysis and feature learning present new research opportunities to large scale multimedia data, e.g. large scale face recognition, intelligent visual surveillance, web-scale image retrieval/annotation /classification, massive object recognition etc. As a consequence, matrix analysis and feature learning are never-ending resilience fields and attract growing efforts from different fields.

This special session hunts for original research results for Matrix Analysis and Feature Learning for Multimedia Understanding. The goals of this special session are twofold: 1) developing matrix analysis and feature learning algorithms to target specific applications in multimedia understanding and large-scale visual data analytics and 2) defining novel large-scale multimedia data driven applications, which can be cleared up by conventional matrix analysis and feature learning algorithms.

Indicative Topics/Areas

Manuscripts are solicited to address a wide range of topics in matrix analysis, feature learning, and multimedia understanding, but not limit to the following:

Extension of traditional matrix analysis and feature learning, multiview dimension reduction, spectral analysis, matrix analysis and feature learning for manifold learning, kernel machines and tensor machines, matrix and tensor analysis for biometrics, matrix and tensor analysis for web-scale multimedia information retrieval, video surveillance, sparse analysis, deep learning, etc.

Important Dates

April 15, 2016: Deadline for submission of full-length papers to special sessions.

May 25, 2016: Acceptance/Rejection Notification.

July 9, 2016: Final camera-ready papers due in electronic form.

Submission

Manuscripts for a Special Session should NOT be submitted in duplication to any other regular or special sessions and should be submitted to SMC 2016 main conference online submission system on SMC 2016 conference website.

All submitted papers of Special Sessions have to undergo the same review process (three completed reviews per paper).  The technical reviewers for each Special Session paper will be members of the SMC 2016 Program Committee and qualified peer-reviewers to be nominated by the Special Session organizers.

Organized by IEEE SMC TC on Cognitive Computing

Special Session organizer

Dr. Xinmei Tian

Professor

University of Science and Technology of China, China

Email:

xinmei@ustc.edu.cn

Co-organizer(s):

Dr. Jun Yu

Associate Professor

Hangzhou Dianzi University, China

Email:

yujun@hdu.edu.cn

Dr. Weifeng Liu

Associate Professor

China University of Petroleum, China

Email:

liuwfxy@gmail.com

Dr. Yicong Zhou

Assistant Professor

University of Macau, Macau, China

E-mail: 

yicongzhou@umac.mo

Submission guidlines:

Special Session: Matrix Analysis and Feature Learning for Multimedia Understanding

A link to upload the special session papers can be found in the webpage as below:

http://140.118.207.10/SMC2016SS/