A Semi-Supervised Learning Algorithm for Growing Neural Gas in Face Recognition

In many classification applications and face recognition tasks, there exist unlabelled data available for training along with labelled samples. The use of unlabelled data can improve the performance of a classifier. In this paper, a semi-supervised growing neural gas is proposed for learning with such partly labelled datasets in face recognition applications. The classifier is first trained on the labelled data and then gradually unlabelled data is classified and added to the training data. The classifier is retrained; and so on. The proposed iterative algorithm conforms to the EM framework and is demonstrated, on both artificial and real datasets, to significantly boost the classification rate with the use of unlabelled data. The improvement is particularly great when the labelled dataset is small. Comparison with support vector machine classifiers is also given. The algorithm is computationally efficient and easy to implement.     

Diagnosis of response and non-response to dry eye treatment using infrared thermography images

The dry eye treatment outcome depends on the assessment of clinical relevance of the treatment effect. The potential approach to assess the clinical relevance of the treatment is to identify the symptoms responders and non-responders to the given treatments using the responder analysis. In our work, we have performed the responder analysis to assess the clinical relevance effect of the dry eye treatments namely, hot towel, EyeGiene®, and Blephasteam® twice daily and 12 min session of Lipiflow®. Thermography is performed at week 0 (baseline), at weeks 4 and 12 after treatment. The clinical parameters such as, change in the clinical irritations scores, tear break up time (TBUT), corneal staining and Schirmer’s symptoms tests values are used to obtain the responders and non-responders groups. We have obtained the infrared thermography images of dry eye symptoms responders and non-responders to the three types of warming treatments. The energy, kurtosis, skewness, mean, standard deviation, and various entropies namely Shannon, Renyi and Kapoor are extracted from responders and non-responders thermograms. The extracted features are ranked based on t-values. These ranked features are fed to the various classifiers to get the highest performance using minimum features. We have used decision tree (DT), K nearest neighbour (KNN), Naves Bayesian (NB) and support vector machine (SVM) to classify the features into responder and non-responder classes. We have obtained an average accuracy of 99.88%, sensitivity of 99.7% and specificity of 100% using KNN classifier using ten-fold cross validation.     

Enhancement of energy efficiency for mechanical production of fine and ultra-fine particles in comminution

This paper presents the energy requirements for mechanical production of fine and ultra-fine particles in comminution. Recent approaches for effective size reduction and energy saving in comminution are described, viz., （a） development and application of new mills/classifiers, （b） adjustment of the bead characterization in stirred bead mills, （c） hybrid comminution systems with roller-press and media mill, （d） assisted methods, and （e） simulation.2007 Chinese Society of Particuology and Institute of Process Engineering, Chinese Academy of Sciences. Published by Elsevier B.V.     

Application of infrared thermography in computer aided diagnosis

The invention of thermography, in the 1950s, posed a formidable problem to the research community: What is the relationship between disease and heat radiation captured with Infrared (IR) cameras? The research community responded with a continuous effort to find this crucial relationship. This effort was aided by advances in processing techniques, improved sensitivity and spatial resolution of thermal sensors. However, despite this progress fundamental issues with this imaging modality still remain. The main problem is that the link between disease and heat radiation is complex and in many cases even non-linear. Furthermore, the change in heat radiation as well as the change in radiation pattern, which indicate disease, is minute. On a technical level, this poses high requirements on image capturing and processing. On a more abstract level, these problems lead to inter-observer variability and on an even more abstract level they lead to a lack of trust in this imaging modality. In this review, we adopt the position that these problems can only be solved through a strict application of scientific principles and objective performance assessment. Computing machinery is inherently objective; this helps us to apply scientific principles in a transparent way and to assess the performance results. As a consequence, we aim to promote thermography based Computer-Aided Diagnosis (CAD) systems. Another benefit of CAD systems comes from the fact that the diagnostic accuracy is linked to the capability of the computing machinery and, in general, computers become ever more potent. We predict that a pervasive application of computers and networking technology in medicine will help us to overcome the shortcomings of any single imaging modality and this will pave the way for integrated health care systems which maximize the quality of patient care.     

Notes on classifying inputs and outputs in data envelopment analysis: A comment

Cook and Zhu (2007) introduced an innovative method to deal with flexible measures. Toloo (2009) found a computational problem in their approach and tackled this issue. Amirteimoori and Emrouznejad (2012) claimed that both Cook and Zhu (2007) and Toloo (2009) models overestimate the efficiency. In this response, we prove that their claim is incorrect and there is no overestimate in these approaches.     

Identifying essential proteins using modified-monkey algorithm (MMA)

Proteins are the most essential macromolecules needed for the normal flow of life. Essential proteins play a key role to control other proteins in an interaction network for the growth and understanding of the molecular mechanism of cellular life. Though there are many computational algorithms for essential drug discovery depending on nature of essential proteins, but still lots of improvements and optimizations are required. In this work, modified-Monkey algorithm (MMA) is proposed for the identification of essential proteins in protein-protein interaction network (PPIN). Nature of a monkey can be distinctly described in three processes like climb, watch-jump, and somersault in different problem spaces. These processes of monkey traversal are plotted in PPIN with objective to find out essential proteins. Performance of MMA is assessed with other existing essential protein prediction methodologies, including Eigenvector Centrality (EC), Betweenness Centrality (BC), Network Centrality (NC) and others also. The proposed methodology has achieved higher success rates in comparison to the existing state-of-art model.     

遥感影像分类方法研究进展

遥感影像分类是遥感信息提取的重要手段,是目前遥感技术中的热点研究内容.分类方法是遥感影像分类的重要内容,有效地选择合适的分类方法是提高分类精度的关键.随着遥感技术的发展,传统的非参数分类方法已经难以满足分类精度需求,基于智能算法的非参数分类方法得到了迅速发展,并在遥感影像分类中发挥着重要作用.近年来,组合分类器由于能够...     

Classifier ensembles for fMRI data analysis: an experiment

Functional magnetic resonance imaging (fMRI) is becoming a forefront brain–computer interface tool. To decipher brain patterns, fast, accurate and reliable classifier methods are needed. The support vector machine (SVM) classifier has been traditionally used. Here we argue that state-of-the-art methods from pattern recognition and machine learning, such as classifier ensembles, offer more accurate classification. This study compares 18 classification methods on a publicly available real data set due to Haxby et al. [Science 293 (2001) 2425–2430]. The data comes from a single-subject experiment, organized in 10 runs where eight classes of stimuli were presented in each run. The comparisons were carried out on voxel subsets of different sizes, selected through seven popular voxel selection methods. We found that, while SVM was robust, accurate and scalable, some classifier ensemble methods demonstrated significantly better performance. The best classifiers were found to be the random subspace ensemble of SVM classifiers, rotation forest and ensembles with random linear and random spherical oracle.