Chungsoo Lim

Oscillometric Blood Pressure Estimation Based on Maximum Amplitude Algorithm Employing Gaussian Mixture Regression

This paper introduces a novel approach to estimate the systolic and diastolic blood pressure ratios (SBPR and DBPR) based on the maximum amplitude algorithm (MAA) using a Gaussian mixture regression (GMR). The relevant features, which clearly discriminate the SBPR and DBPR according to the targeted groups, are selected in a feature vector. The selected feature vector is then represented by the Gaussian mixture model. The SBPR and DBPR are subsequently obtained with the help of the GMR and then mapped back to SBP and DBP values that are more accurate than those obtained with the conventional MAA method.

A new a priori SNR estimator based on multiple linear regression technique for speech enhancement

Article history: Available online 18 April 2014 We propose a new approach to estimate the ap riori signal-to-noise ratio (SNR) based on a multiple linear regression (MLR) technique. In contrast to estimation of the ap riori SNR employing the decision- directed (DD) method, which uses the estimated speech spectrum in previous frame, we propose to find the ap riori SNR based on the MLR technique by incorporating regression parameters such as the ratio between the local energy of the noisy speech and its derived minimum along with the a posteriori SNR. In the experimental step, regression coefficients obtained using the MLR are assigned according to various noise types, for which we employ a real-time noise classification scheme based on a Gaussian mixture model (GMM). Evaluations using both objective speech quality measures and subjective listening tests under various ambient noise environments show that the performance of the proposed algorithm is better than that of the conventional methods.

Efficient implementation techniques of an SVM-based speech/music classifier in SMV

For real-time speech and audio encoders used in various multimedia applications, low-complexity encoding algorithms are required. Indeed, accurate classification of input signals is the key prerequisite for variable bit rate encoding, which has been introduced in order to effectively utilize limited communication bandwidth. This paper investigates implementation issues with a support vector machine (SVM)-based speech/music classifier in the selectable mode vocoder (SMV) framework, which is a standard codec adopted by the Third-Generation Partnership Project 2 (3GPP2). While a support vector machine is well known for its superior classification capability, it is accompanied by a high computational cost. In order to achieve a more realizable system, we propose two techniques for the SVM-based speech/music classifier, aimed at reducing the number of classification requests to the classifier. The first technique introduces a simpler classifier that processes some of the input frames instead of the SVM-based classifier, and the second technique skips a portion of input frames based on strong inter-frame correlation in speech and music frames. Our experimental results show that the proposed techniques can reduce the computational cost of the SVM-based classifier by 95.4 % with negligible performance degradation, making it plausible for integration into the SMV codec.

Efficient implementation of an SVM-based speech/music classifier by enhancing temporal locality in support vector references

Speech/music classification is an integral part of various consumer electronics applications such as audio codecs, multimedia document indexing, and automatic speech recognition. To achieve high performance at speech/music classification, a support vector machine (SVM) has been widely used as a classifier due to its decent classification capability. However, in order to use an SVM-based speech/music classifier in embedded systems, which gradually replace desktop computer systems, one significant implementation problem needs to be resolved: high implementation cost due to time and energy inefficiency. The memory requirement determined by the dimensionality and the number of support vectors, is generally too high for an embedded systems cache to accommodate resulting in expensive memory accesses. In this paper, two techniques are proposed to reduce expensive memory accesses by enhancing temporal locality in support vector references utilizing fetched data from memory with great efficiency. For this, the patterns in support vector references are first analyzed, and then loop transformation techniques are proposed to improve the temporal locality that register file and cache hierarchy take advantage of. The proposed techniques are evaluated by applying them to a speech codec, and the enhancement is confirmed by measuring the number of memory accesses, overall execution time, and energy consumption.

A new television audience measurement framework using smart devices

Television audience measurement is intended to collect information on the audiences watching a specific television program at a particular time. This information is crucial for television broadcasters and advertisers because they need to provide right television programs and commercials to right audiences to maximize their investments in broadcasting. For accurate measurements, a panel of representative audiences must be selected judiciously so that it accurately represents the entire target audience group. However, it is hard to secure a proper number of target audiences due to the expensive and cumbersome installations of measurement equipments. To resolve this issue in panel selection, we propose a novel television audience measurement framework using pervasive smart devices such as a smartphone. In the proposed framework, a short audio signal from a television set is recorded by a personal smart device and is sent to an audio matching server for the identification of the television program shown by the television set. For effective identification, we propose an accurate audio matching algorithm based on spectral coherence and efficient implementation techniques that exploit the inherent parallelism in the algorithm. To verify the plausibility of the framework and the effectiveness of the audio matching algorithm, we conduct experiments with diverse genres of television programs under various recording conditions.

New techniques for improving the practicality of an SVM-based speech/music classifier

Variable bit-rate coding introduced for effective utilization of limited communication bandwidth requires accurate classification of input signals. This paper investigates implementation of a support vector machine (SVM)-based speech/music classifier in the selectable mode vocoder (SMV) framework, which is a standard codec adopted by the Third-Generation Partnership Project 2 (3GPP2). A support vector machine is well known for its superior pattern recognition capability; however, it is accompanied by a high computational cost. In order to achieve a more practical system, three techniques are proposed for the SVM-based speech/music classifier. The first is to prune support vectors that least contribute to the output of the SVM, while the other two are aimed at reducing the number of classification requests to the SVM-based classifier by eliminating or redirecting some of the classification requests to the classifier.

Improving SVM-based speech/music classifier based on conditional MAP criterion

To enhance SVM-based speech/music classifications by exploiting substantial inter-frame correlations in speech/music frames, we propose a technique using the second-order conditional maximum a posteriori (CMAP) probability that considers not only the current observation but also the classification results of the two previous frames.

Efficient Implementation of SVM-Based Speech/Music Classification on Embedded Systems

Accurate classification of input signals is the key prerequisite for variable bit-rate coding, which has been introduced in order to effectively utilize limited communication bandwidth. Especially, recent surge of multimedia services elevate the importance of speech/music classification. Among many speech/music classifier, the ones based on support vector machine (SVM) have a strong selling point, high classification accuracy, but their computational complexity and memory requirement hinder their way into actual implementations. Therefore, techniques that reduce the computational complexity and the memory requirement is inevitable, particularly for embedded systems. We first analyze implementation of an SVM-based classifier on embedded systems in terms of execution time and energy consumption, and then propose two techniques that alleviate the implementation requirements: One is a technique that removes support vectors that have insignificant contribution to the final classification, and the other is to skip processing some of input signals by virtue of strong correlations in speech/music frames. These are post-processing techniques that can work with any other optimization techniques applied during the training phase of SVM. With experiments, we validate the proposed algorithms from the perspectives of classification accuracy, execution time, and energy consumption.