2014年10月18日，课题组成员李阳、张连波参加了在武汉召开的2014年安全、模式分析与控制论国际会议（ICSPAC2014），并对各自的录用论文进行了报告。ICSPAC2014（2014年安全、模式分析与控制论国际会议）是由华中科技大学与澳门大学联合承办的国际会议，武汉市科学技术协会和IEEE Systems, Man and Cybernetics Society为大会提供了赞助。该会议在湖北武汉举行，提供了一个国际性的学术论坛，其中汇集信息科学和工程，安全控制，模式识别与分析方面，报告了最新创意和发展状况，并总结了最先进的发展成果，同时在会场就安全模式分析和控制科学方面的进步发展做了思想交流。

本次ICSPAC2014会议，课题组共有4篇论文被录用。

1）Co-Regularization for Classification
Yang Li, Dapeng Tao, Weifeng Liu*,Yanjiang Wang
Abstract—Semi-supervised learning algorithms that combine labeled and unlabeled data receive significant interests in recent years and are successfully deployed in many practical data mining applications. Manifold regularization, one of the most representative works, tries to explore the geometry of the intrinsic data probability distribution by penalizing the classification function along the implicit manifold. Although existing manifold regularization, including Laplacian regularization (LR) and Hessian regularization (HR), yields significant benefits for partially labeled classification, it is observed that LR suffers from the poor generalization and HR exhibits the characteristic of instability, both manifold regularization could not accurately reflect the ground-truth. To remedy the problems in single manifold regularization and approximate the intrinsic manifold, we propose Manifold Regularized Co-Training(Co-Re) framework， which combines the manifold regularization (LR and HR) and the algorithm co-training. Extensive experiments on the USAA video dataset are conducted and validate the effectiveness of Co-Re by comparing it with baseline manifold regularization algorithms.
Keywords—semi-supervised learning; manifold regularization; Co-Training; Laplacian regularization; Hessian regularization

2）Laplacian Regularized Active Learning for Image Segmentation
Lianbo Zhang, Dapeng Tao, Weifeng Liu* 
Abstract—Image segmentation is a common topic in image processing. Many methods has been used in image segmentation, such as Graph cut, threshold-based. However, these methods can’t work with high precision. Among these method,  SVM is used as a good tool for classification, as we treat image segmentation as a problem of classification. To solve the problem above and get better segmentation result as well as high precision, we add Laplacian regularization to SVM algorithm to get a new algorithm i.e. Laplacian regularized active learning for image segmentation. Our algorithm considers distance between pixels when segmenting a picture, which is executed by Laplacian regularization. Experiments demonstrate that our algorithm perform better in comparison with common SVM algorithm.
Keywords—laplacian; active learning; image segmentation; support vector machine

3）Laplacian-Hessian regularization for Semi-supervised Classification
Hongli Liu, Weifeng Liu, Dapeng Tao, Yanjiang Wang

Abstract—With exploiting a small number of labeled images and a large number of unlabeled images, semi-supervised learning has attracted centralized attention in recent years. The representative works are Laplacian and Hessian regularization methods. However, Laplacian method tends to a constant value and poor generalization in the process of classification. Although Hessian energy can properly forecast the data points beyond the range of the domain, its regularizer probably leads to useless results in the process of regression. So the Laplacian-Hessian regression method for image classification is proposed, which can both predict the data points and enhance the stability of Hessian regularizer. To evaluate the Laplacian-Hessian method, Columbia Consumer Video database is employed in the paper. Experimental results demonstrate that the proposed method perform better than Laplacian or Hessian method in the matter of classification and stability. 

Keywords—Laplacian; Hessian; Stability; semi-supervised learning

4）Design of general-purpose acquisition-control module for well-logging signal based on DSP and FPGA

Xu Dahua, Wang Jun

Abstract—The signal-acquisition-control-modules of existing well-logging tools have poor universality, low integration and poor stability. A general-purpose signal-acquisition-control module for versatile well-logging tool is designed which is based on SM320F28335 DSP chip and A3P250 FPGA chip. The composition of the hardware circuit is given. The acquisition on analog signals, pulse signals, as well as waves can be completed by this module, which can control both relay and analog switch. The module can provide versatile bus interfaces. On different well-logging tools, no change is needed for the hardware circuit, what’s only needed is to download the corresponding procedure. Tests show that this module works stably and has high performance and low failure rate, which shows that it is suitable for the mal-condition of high temperature and pressure underground.
Keywords—DSP FPGA; well-logging tool; general-purpose; signal-acquisition-control-module