Self-splitting Competitive Learning Research Articles

An approach for complex activity recognition by key frames

A new method for complex activity recognition in videos by key frames was presented. The progressive bisection strategy (PBS) was employed to divide the complex activity into a series of simple activities and the key frames representing the simple activities were extracted by the self-splitting competitive learning (SSCL) algorithm. A new similarity criterion of complex activities was defined. Besides the regular visual factor, the order factor and the interference factor measuring the timing matching relationship of the simple activities and the discontinuous matching relationship of the simple activities respectively were considered. On these bases, the complex human activity recognition could be achieved by calculating their similarities. The recognition error was reduced compared with other methods when ignoring the recognition of simple activities. The proposed method was tested and evaluated on the self-built broadcast gymnastic database and the dancing database. The experimental results prove the superior efficiency.

Data clustering algorithm for channel segmentation in a radio monitoring system

The detection of signals and the estimation of signal bandwidth is a perpetual topic in radio communication systems. Both issues are extremely challenging, since the wireless channel is unreliable in nature. A radio monitoring system faces the most difficult conditions in this task; it normally scans a wide frequency range of several hundred MHz and has to detect a multitude of different signals. Owing to the computational costs, the radio monitoring systems use nowadays mainly energy detectors based on fast Fourier transform spectrum analysers and a static threshold, defined by a previous noise estimation. A refined algorithm based on the self-splitting competitive learning (SSCL) clustering is presented that quantises the power spectral density (PSD) according to the present signal power levels. The quantisation of the PSD results in a promising channel segmentation. In contrast to the traditional threshold evaluation, this approach is independent of a previously assumed noise estimation and therefore more robust against noise level and noise distribution changes. The presented definition of the essential cluster validity criterion is key for a successful channel segmentation. Furthermore, the novel postprocessing of the clustering result introduced in this study evaluates the progression of the PSD data and significantly improves the channel segmentation.

Exploitation of source nonstationarity in underdetermined blind source separation with advanced clustering techniques

The problem of blind source separation (BSS) is investigated. Following the assumption that the time-frequency (TF) distributions of the input sources do not overlap, quadratic TF representation is used to exploit the sparsity of the statistically nonstationary sources. However, separation performance is shown to be limited by the selection of a certain threshold in classifying the eigenvectors of the TF matrices drawn from the observation mixtures. Two methods are, therefore, proposed based on recently introduced advanced clustering techniques, namely Gap statistics and self-splitting competitive learning (SSCL), to mitigate the problem of eigenvector classification. The novel integration of these two approaches successfully overcomes the problem of artificial sources induced by insufficient knowledge of the threshold and enables automatic determination of the number of active sources over the observation. The separation performance is thereby greatly improved. Practical consequences of violating the TF orthogonality assumption in the current approach are also studied, which motivates the proposal of a new solution robust to violation of orthogonality. In this new method, the TF plane is partitioned into appropriate blocks and source separation is thereby carried out in a block-by-block manner. Numerical experiments with linear chirp signals and Gaussian minimum shift keying (GMSK) signals are included which support the improved performance of the proposed approaches.

Self-splitting competitive learning: a new on-line clustering paradigm

Clustering in the neural-network literature is generally based on the competitive learning paradigm. The paper addresses two major issues associated with conventional competitive learning, namely, sensitivity to initialization and difficulty in determining the number of prototypes. In general, selecting the appropriate number of prototypes is a difficult task, as we do not usually know the number of clusters in the input data a priori. It is therefore desirable to develop an algorithm that has no dependency on the initial prototype locations and is able to adaptively generate prototypes to fit the input data patterns. We present a new, more powerful competitive learning algorithm, self-splitting competitive learning (SSCL), that is able to find the natural number of clusters based on the one-prototype-take-one-cluster (OPTOC) paradigm and a self-splitting validity measure. It starts with a single prototype randomly initialized in the feature space and splits adaptively during the learning process until all clusters are found; each cluster is associated with a prototype at its center. We have conducted extensive experiments to demonstrate the effectiveness of the SSCL algorithm. The results show that SSCL has the desired ability for a variety of applications, including unsupervised classification, curve detection, and image segmentation.