A New Distance Measure for a Variable-Sized Acoustic Model Based on MDL Technique

This paper presents a new technique for region-oriented video compression based on the minimization of a minimum description length (MDL) criterion. The proposed scheme is based on a region-based motion compensation approach, in which the displaced frame difference (DFD) is coded together with the segmentation map and the motion parameters. The region-based segmentation is approximated by a polygonal representation which permits us to significantly reduce the segmentation coding cost. The polygonal approximation, which minimizes the MDL criterion, is considered the optimal approximation in terms of compression performance. In this contest, the MDL criterion has been developed and tested for the three following DFD coding techniques: (1) entropy coding of the quantized DFD image; (2) entropy coding of the quantized DFD values obtained after morphological filtering of the DFD image; and (3) DCT-based coding of the DFD image. The experimental results, presented on real video sequences, validate the proposed approach.

An approach for using minimum description length (MDL) criterion for autoregressive moving-average (ARMA) models and autoregressive models with exogenous input (ARX) is formulated in terms of the minimum eigenvalue of a covariance matrix. It is shown that as the number of observation data points increases without bound, the MDL criterion reduces to a simple monotonic nonlinear transformation of this minimum eigenvalue. Consequently, it is shown that in the asymptotic case the minimum-eigenvalue provides exactly the same information as the original MDL formulation. Additionally, the approach does not require prior estimation of the model parameters, and thus has the potential for significantly reduced computation compared to other MDL-based model order determination schemes. The approach's application to the selection of ARMA and ARX model orders is discussed, and several numerical examples are presented. >

Outlier detection is imperative in biomedical data analysis to achieve reliable knowledge discovery. In this paper, a new outlier detection method based on Kullback-Leibler (KL) divergence is presented. The original concept of KL divergence was designed as a measure of distance between two distributions. Stemming from that, we extend it to biological sample outlier detection by forming sample sets composed of nearest neighbors. To handle the non-linearity during the KL divergence calculation and to tackle with the singularity problem due to small sample size, we map the original data into a higher feature space and apply kernel functions without resorting to a mapping function. A sample possessing the largest KL divergence is detected as an outlier. The proposed method is tested with one synthetic data, two public gene expression data sets, and our own mass spectrometry data generated for prostate cancer study.

A novel method for enhancement of noisy speech is presented. Frames of speech samples are split into low and high frequency bands and projected onto a library of bases consisting of local trigonometric functions and wavelet packets. Coefficients thought to represent only the speech are selected by means of the minimum description length (MDL) criterion, and used to synthesize an estimate of the original speech. A tracking algorithm uses MDL values to choose between the MDL processor and alternate processors which reject audible artifacts. Preliminary results indicate that the new algorithm may be useful in applications requiring a single-microphone noise reduction system for speech.

We present a novel approach for clip-based key frame extraction. Our framework allows both clips with subtle changes as well as clips containing rapid shot changes, fades and dissolves to be well approximated. We show that creating key frame video abstractions can be achieved by transforming each frame of a video sequence into an eigenspace and then clustering this space using Gaussian mixture models (GMMs). A minimum description length (MDL) criterion is then used to determine the optimal number of GMM components to use in the clustering. The image nearest to the centres of each of the GMM components are selected as key frames. Unlike previous work, this technique relies on global video clip properties and results show that the key frames extracted give a very good representation of the overall clip content. We demonstrate the application of this technique on a database of 307 clips of wildlife footage containing dissolves, shot changes, fades, pans, zooms and a wide range of animal behaviours.

In this paper, we employ the Kullback-Leibler Divergence (KD) measure from information theory within the framework of Self Organizing Maps applied to network planning and dimensioning. Here, the KD measure estimates the closeness among probability densities (or relative Entropy), and is used instead of the distance or dot product measures. Applied to a real network planning under the supervision of link budget calculations, the new method shows improved performance and a promise of wider network applications.

In this paper we present an experimental study exploiting structural Bayesian adaptation for handling potential mismatches between training and test conditions for real-world applications to be realized in our multilingual very large vocabulary speech recognition (VLVSR) system project sponsored by MOTIE (The Ministry of Trade, Industry and Energy), Republic of Korea. The goal of the project is to construct a national-wide VLVSR cloud service platform for mobile applications. Besides system architecture design issues, at such a large scale, performance robustness problems, caused by mismatches in speakers, tasks, environments, and domains, etc., need to be taken into account very carefully as well. We decide to adopt adaptation, especially the structural MAP, techniques to reduce system accuracy degradation caused by these mismatches. Being part of an ongoing project, we describe how structural MAP approaches can be used for adaptation of both acoustic and language models for our VLVSR systems, and provide convincing experimental results to demonstrate how adaptation can be utilized to bridge the performance gap between the current state-of-the-art and deployable VLVSR systems.

An minimum description length (MDL) criterion is proposed to choose a good partition for a bipartite network. A heuristic algorithm based on combination theory is presented to approach the optimal partition. As the heuristic algorithm automatically searches for the number of partitions, no user intervention is required. Finally, experiments are conducted on various datasets, and the results show that our method generates higher quality results than the state-of-art methods, cross-association and bipartite, recursively induced modules. Experiment results also show the good scalability of the proposed algorithm. The method is applied to traditional Chinese medicine (TCM) formula and Chinese herbal network whose community structure is not well known, and found that it detects significant and it is informative community division.

We present a novel approach for prediction of satellite image frame that uses a fuzzy rule based framework. The input-output membership functions for the premise and consequent parts of the rules are derived using a Gaussian Mixture Model (GMM). The weights of the fuzzy rules are represented as the prior probabilities of the respective Gaussian components. For obtaining the predictive fuzzy model, the GMM parameters are estimated via EM algorithm using a spatiotemporal representation of image sequence or video clips. Minimum Description Length (MDL) criterion is used to obtain a suitable predictive fuzzy model. The resulting model is successfully applied on a sequence of satellite images of tropical cyclone, Nargis, that made landfall in Myanmar on May 2, 2008. The quality of the predicted image is assessed using two criteria. The proposed approach is found to predict image frame successfully.

This paper presents a new technique for region-oriented video compression based on the minimization of a minimum description length (MDL) criterion. The proposed video compression scheme is based on a region-based motion compensation approach in which the displaced frame difference is partially coded together with the segmentation and the motion parameters. The region-based segmentation is approximated by a polygonal representation which permits to significantly reduce the segmentation coding cost. The polygonal approximation which minimizes the MDL criterion is finally considered as optimal. Experimental results are presented on real video sequences.

In this paper we propose an expectation-maximization (EM) algorithm with distance measure for color image segmentation. The probability distribution model used is the Gaussian mixture model. The concept of color distance measure is used in this algorithm to determine the region to which a particular pixel belongs. L *a* b color space is used to replace the more straightforward spaces such as the RGB color space and YUV color space. This algorithm is capable of automatically selecting the number of components of the model using minimum description length (MDL) criterion. The proposed method yields good segmentation with better PSNR and SSIM values compared to classical EM algorithm; that is, the segmented image will be structurally more similar to the original image.

This letter proposes a channel prediction scheme for precoded time division duplex multiple-input multiple-output (TDD-MIMO) systems. Based on the sum-of-sinusoidal (SoS) channel model, the proposed scheme first uses iterative adaptive approach (IAA)-based spectral estimation algorithm for sinusoidal components estimation according to the channel state information (CSI) detected in the discontinuous uplink periods. Then the minimum description length (MDL) criterion is used to choose the proper number of sinusoidal components for the channel model. Finally, time extrapolation is performed to the estimated channel model to predict the future downlink CSI that is needed for transmit precoding. The scheme is verified by simulations.

A hybrid model comprised of Gaussian Mixtures Models (GMMs) and Hidden Markov Models (HMMs) is used to model generic sounds with large intra class perceptual variations. Each class has variable number of mixture components in the GMM. The number of mixture components is derived using the Minimum Description Length (MDL) criterion. The overall performance of the hybrid model was compared against models based on HMMs and GMMs with a fixed number of mixture components across all classes. We show that a hybrid model outperforms both class-based GMMs, HMMs, and GMMs based on fixed number of components. Further, our experiments revealed that the contribution of transitions between states in HMMs has no significant effect on the overall classification performance of generic sounds when large intra class perceptual variations are present among sounds in the training and test datasets. Sounds that show multi-event structure with events that tend to be similar (repetitive) indicated improved performance when modeled with HMMs that can be attributed to HMM’s state transition property. Conversely, GMMs indicate better performance when the sound samples show subtle or no repetitive behavior. These results were validated using the MuscleFish sound database.

In this paper we have estimated signal to noise ratio (SNR) of active users in uplink scenario of CDMA systems blindly. We have not any knowledge about PN (pseudo noise) code, frequency, and timing of received signals. In this way first we synchronized received signal of desired user and then we use subspace method for SNR estimation of synchronized user. This type of SNR estimation is classified as non data aided (NDA) methods. Our synchronization is based on maximizing of covariance matrix Frobenius norm. After that by estimating PN code of strongest signal we are able to estimate transmitted signal data, and by using estimated SNR, we can eliminate this signal from received signal by using SIC (serial interference canceller). By repetition this cycle for each user we can estimate SNR of active users in uplink scenario. Also, in this paper we use minimum description length (MDL) criteria for estimating signal subspace dimension. Standard divisions of estimated SNR show that good agreement with exact SNR for all users.

Network-assisted diversity multiple access (NDMA) was recently proposed to resolve collision. The blind NDMA method using independent component analysis (ICA) ICA-NDMA, has been developed to overcome the difficulties of synchronization and orthogonal identification codes required by the training-based NDMA protocols. In this paper, we propose a novel blind collision resolution protocol, named as ICA-CoopNDMA, by employing cooperative transmission in ICA-NDMA. When K-node collision happened, a cooperative transmission mechanism is adopted to collect different mixtures of the K original packets in the following slots. After that, the destination can recover the original packets using ICA. This protocol offers spatial diversity by using relay nodes and is robust to the wireless channel. In order to determine the collision multiplicity K, a cooperative blind detection algorithm is also proposed based on minimum description length (MDL) criterion. This algorithm saves one detection slot compared with conventional MDL-based method used by ICA-NDMA. Hence, the delay of ICA-CoopNDMA is reduced. Simulation and theoretical analysis results show that, the proposed protocol can work effectively under both fast and slow fading channels. Compared with ICA-NDMA, the maximum stable throughput and average packet delay are obviously better, and the energy efficiency is higher under heavy traffic load.