This paper presents an unsupervised learning-based method for selection of feature points and object category classification without previous setting of the number of categories. Our method consists of the following procedures: 1)detection of feature points and description of features using a Scale-Invariant Feature Transform (SIFT), 2)selection of target feature points using One Class-Support Vector Machines (OC-SVMs), 3)generation of visual words of all SIFT descriptors and histograms in each image of selected feature points using Self-Organizing Maps (SOMs), 4)formation of labels using Adaptive Resonance Theory-2 (ART-2), and 5)creation and classification of categories on a category map of Counter Propagation Networks (CPNs) for visualizing spatial relations between categories. Classification results of static images using a Caltech-256 object category dataset and dynamic images using time-series images obtained using a robot according to movements respectively demonstrate that our method can visualize spatial relations of categories while maintaining time-series characteristics. Moreover, we emphasize the effectiveness of our method for category classification of appearance changes of objects.

This study presents three image processing systems for snow particle classification into snowflake and graupel. All of them are based on feature classification, yet as a novelty in all cases multiple features are exploited. Additionally, each of them is characterized by a different data flow. In order to compare the performances, we not only consider various features, but also suggest different classifiers. The best achieved results are for the snowflake discrimination method applied before statistical classifier, as the correct classification ratio in this case reaches 94%. In other cases the best results are around 88%.

The traditional mobile consumer electronics such as media players and smart phones use two distinct memories, SDRAM and Flash memory. SDRAM is used as main memory since it has characteristic of byte-unit random accessibility while Flash memory as secondary storage due to its characteristic of non-volatility. However, the advent of Storage Class Memory (SCM) that supports both SDRAM and Flash memory characteristics gives an opportunity to design a new system configuration. In this paper, we explore four feasible system configurations, namely RAM-Flash, RAM-SCM, SCM-Flash and SCM-Only. Then, using a real embedded system equipped with FeRAM, a type of SCM, we analyze the tradeoffs between performance and energy efficiency of each configuration. Experimental results have shown that SCM has great potential to reduce energy consumption for all configurations while performance is highly application dependent and might be degraded on the SCM-Flash and SCM-Only configuration.

A novel approach for fast traffic classification for the high speed networks is proposed, which bases on the protocol behavior statistical features. The packet size and a new parameter named "Estimated Protocol Processing Time" are collected from the real data flows. Then a set of joint probability distributions is obtained to describe the protocol behaviors and classify the traffic. Comparing the parameters of an unknown flow with the pre-obtained joint distributions, we can judge which application protocol the unknown flow belongs to. Distinct from other methods based on traditional inter-arrival time, we use the "Estimated Protocol Processing Time" to reduce the location dependence and time dependence and obtain better results than traditional traffic classification method. Since there is no need for character string searching and parallel feature for hardware implementation with pipeline-mode data processing, the proposed approach can be easily deployed in the hardware for real-time classification in the high speed networks.

Kernel logistic regression (KLR) is a powerful and flexible classification algorithm, which possesses an ability to provide the confidence of class prediction. However, its training--typically carried out by (quasi-)Newton methods--is rather time-consuming. In this paper, we propose an alternative probabilistic classification algorithm called Least-Squares Probabilistic Classifier (LSPC). KLR models the class-posterior probability by the log-linear combination of kernel functions and its parameters are learned by (regularized) maximum likelihood. In contrast, LSPC employs the linear combination of kernel functions and its parameters are learned by regularized least-squares fitting of the true class-posterior probability. Thanks to this linear regularized least-squares formulation, the solution of LSPC can be computed analytically just by solving a regularized system of linear equations in a class-wise manner. Thus LSPC is computationally very efficient and numerically stable. Through experiments, we show that the computation time of LSPC is faster than that of KLR by two orders of magnitude, with comparable classification accuracy.

This paper presents a combined feature extraction method to improve the performance of bag-of-features image classification. We apply 10 relevant operations to global/local statistics of visual words. Because the pairwise combination of visual words is large, we apply feature selection methods including fisher discriminant criterion and L1-SVM. The effectiveness of the proposed method is confirmed through the experiment.

In this paper, we propose a new speaker-class modeling and its adaptation method for the LVCSR system and evaluate the method on the Corpus of Spontaneous Japanese (CSJ). In this method, closer speakers are selected from training speakers and the acoustic models are trained by using their utterances for each evaluation speaker. One of the major issues of the speaker-class model is determining the selection range of speakers. In order to solve the problem, several models which have a variety of speaker range are prepared for each evaluation speaker in advance, and the most proper model is selected on a likelihood basis in the recognition step. In addition, we improved the recognition performance using unsupervised speaker adaptation with the speaker-class models. In the recognition experiments, a significant improvement could be obtained by using the proposed speaker adaptation based on speaker-class models compared with the conventional adaptation method.

This paper proposes a self-taught classifier of gateways for hybrid SLAM. Gateways are detected and recognized by the self-taught classifier, which is a SVM classifier and self-taught in that its training samples are produced and labeled without user's intervention. Since the detection of gateways at the topological boundaries of an acquired metric map reduces computational complexity in partitioning the metric map into sub-maps as compared with previous hybrid SLAM approaches using spectral clustering methods, from O(2n) to O(n), where n is the number of sub-maps. This makes possible real time hybrid SLAM even for large-scale metric maps. We have confirmed that the self-taught classifier provides satisfactory consistency and computationally efficiency in hybrid SLAM through different experiments.

In this paper, we present a novel color independent components based SIFT descriptor (termed CIC-SIFT) for object/scene classification. We first learn an efficient color transformation matrix based on independent component analysis (ICA), which is adaptive to each category in a database. The ICA-based color transformation can enhance contrast between the objects and the background in an image. Then we compute CIC-SIFT descriptors over all three transformed color independent components. Since the ICA-based color transformation can boost the objects and suppress the background, the proposed CIC-SIFT can extract more effective and discriminative local features for object/scene classification. The comparison is performed among seven SIFT descriptors, and the experimental classification results show that our proposed CIC-SIFT is superior to other conventional SIFT descriptors.

In this paper, we present the sexual dimorphism analysis in 3D human face and perform gender classification based on the result of sexual dimorphism analysis. Four types of features are extracted from a 3D human-face image. By using statistical methods, the existence of sexual dimorphism is demonstrated in 3D human face based on these features. The contributions of each feature to sexual dimorphism are quantified according to a novel criterion. The best gender classification rate is 94% by using SVMs and Matcher Weighting fusion method. This research adds to the knowledge of 3D faces in sexual dimorphism and affords a foundation that could be used to distinguish between male and female in 3D faces.

A low-voltage class-AB CMOS output stage with a tunable quiescent current control circuit is presented. It is based on a complementary common source. The quiescent current is detected by a compact circuit and can be adjusted by means of a control current without need to modify the transistor dimensions. The minimum supply voltage can be down to one threshold voltage plus two saturation voltages. It is suitable to drive low resistive loads. Simulation results are provided that are in agreement with expected characteristics.

A scheme for designing a hierarchical fuzzy classification system with a different number of fuzzy partitions based on statistical characteristics of the data is proposed. To minimize the number of misclassified patterns in intermediate layers, a method of fuzzy partitioning from the defuzzified outputs of previous layers is also presented. The effectiveness of the proposed scheme is demonstrated by comparing the results from five datasets in the UCI Machine Learning Repository.

This paper presents a novel algorithm for Nearest Neighbor (NN) classifier. NN classification is a well-known method of pattern classification having the following properties: * it performs maximum-margin classification and achieves less than twice the ideal Bayesian error, * it does not require knowledge of pattern distributions, kernel functions or base classifiers, and * it can naturally be applied to multiclass classification problems. Among the drawbacks are A) inefficient memory use and B) ineffective pattern classification speed. This paper deals with the problems A and B. In most cases, NN search algorithms, such as k-d tree, are employed as a pattern search engine of the NN classifier. However, NN classification does not always require the NN search. Based on this idea, we propose a novel algorithm named k-d decision tree (KDDT). Since KDDT uses Voronoi-condensed prototypes, it consumes less memory than naive NN classifiers. We have confirmed that KDDT is much faster than NN search-based classifier through a comparative experiment (from 9 to 369 times faster than NN search based classifier). Furthermore, in order to extend applicability of the KDDT algorithm to high-dimensional NN classification, we modified it by incorporating Gabriel editing or RNG editing instead of Voronoi condensing. Through experiments using simulated and real data, we have confirmed the modified KDDT algorithms are superior to the original one.

In this paper, an ultra low power analog front-end for EPCglobal Class 1 Generation 2 RFID tag is presented. The proposed RFID tag removes the need for high frequency clock and counters used in conventional tags, which are the most power hungry blocks. The proposed clock-free decoder employs an analog integrator with an adaptive current source that provides a uniform decoding margin regardless of the data rate and a link frequency extractor based on a relaxation oscillator that generates frequency used for backscattering. A dual supply voltage scheme is also employed to increase the power efficiency of the tag. In order to improve the tolerance of the proposed circuit to environmental variations, a self-calibration circuit is proposed. The proposed RFID analog front-end circuit is designed and simulated in 0.25 µm CMOS, which shows that the power consumption is reduced by an order magnitude compared to the conventional RFID tags, without losing immunity to environmental variations.

Bag-of-Visual-Words representation has recently become popular for scene classification. However, learning the visual words in an unsupervised manner suffers from the problem when faced these patches with similar appearances corresponding to distinct semantic concepts. This paper proposes a novel supervised learning framework, which aims at taking full advantage of label information to address the problem. Specifically, the Gaussian Mixture Modeling (GMM) is firstly applied to obtain "semantic interpretation" of patches using scene labels. Each scene induces a probability density on the low-level visual features space, and patches are represented as vectors of posterior scene semantic concepts probabilities. And then the Information Bottleneck (IB) algorithm is introduce to cluster the patches into "visual words" via a supervised manner, from the perspective of semantic interpretations. Such operation can maximize the semantic information of the visual words. Once obtained the visual words, the appearing frequency of the corresponding visual words in a given image forms a histogram, which can be subsequently used in the scene categorization task via the Support Vector Machine (SVM) classifier. Experiments on a challenging dataset show that the proposed visual words better perform scene classification task than most existing methods.

For an arbitrary set B of Boolean functions satisfying a certain condition, we give a general method of constructing a class CB of read-once Boolean formulas over the basis B that has the following property: For any F in CB, F can be transformed to an optimal formula (i.e., a simplest formula over the standard basis {AND, OR, NOT}) by replacing each occurrence of a basis function h ∈ B in F with an optimal formula for h. For a particular set of basis functions B* = {AND,OR,NOT,XOR,MUX}, we give a canonical form representation for CB* so that the set of canonical form formulas consists of only NPN-representatives in CB*.

In differentiated services, packet classification is used to categorize incoming packets into multiple forwarding classes based on pre-defined filters and make information accessible for quality of service. Although numerous algorithms have presented novel data structures to improve the search performance of packet classification, the performance of these algorithms are usually limited by the characteristics of filter databases. In this paper, we use a different approach of filter preprocessing to enhance the search performance of packet classification. Before generating the searchable data structures, we cluster filters in a bottom-up manner. The procedure of the filter clustering merges filters with high degrees of similarity. The experimental results show that the technique of filter clustering could significantly improve the search performance of Pruned Tuple Space Search, a notable hash-based algorithm. As compared to the prominent existing algorithms, our enhanced Pruned Tuple Space Search also has superior performance in terms of speed and space.

In the last decade, the technique of packet classification has been widely deployed in various network devices, including routers, firewalls and network intrusion detection systems. In this work, we improve the performance of packet classification by using multiple hash tables. The existing hash-based algorithms have superior scalability with respect to the required space; however, their search performance may not be comparable to other algorithms. To improve the search performance, we propose a tuple reordering algorithm to minimize the number of accessed hash tables with the aid of bitmaps. We also use pre-computation to ensure the accuracy of our search procedure. Performance evaluation based on both real and synthetic filter databases shows that our scheme is effective and scalable and the pre-computation cost is moderate.

In this paper, we propose a robust room inverse filtering algorithm for speech dereverberation based on a kurtosis maximization. The proposed algorithm utilizes a new normalized kurtosis function that nonlinearly maps the input kurtosis onto a finite range from zero to one, which results in a kurtosis warping. Due to the kurtosis warping, the proposed algorithm provides more stable convergence and, in turn, better performance than the conventional algorithm. Experimental results are presented to confirm the robustness of the proposed algorithm.

We exhibit a simple procedure to find how classical signals should be processed in cluster-state quantum computation. Using stabilizers characterizing a cluster state, we can easily find a precise classical signal-flow that is required in performing cluster-state computation.