• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.7
• /
• pp.1307-1311
• /
• 2006

In this paper, we propose new speech feature parameter using the Matrix Factorization for appearance part-based features of speech spectrum. The proposed parameter represents effective dimensional reduced data from multi-dimensional feature data through matrix factorization procedure under all of the matrix elements are the non-negative constraint. Reduced feature data presents p art-based features of input data. We verify about usefulness of NMF(Non-Negative Matrix Factorization) algorithm for speech feature extraction applying feature parameter that is got using NMF in Mel-scaled filter bank output. According to recognition experiment results, we confirm that proposed feature parameter is superior to MFCC(Mel-Frequency Cepstral Coefficient) in recognition performance that is used generally.

• ETRI Journal
• /
• v.39 no.1
• /
• pp.21-29
• /
• 2017

This paper addresses the problem of unsupervised speech separation based on robust non-negative matrix factorization (RNMF) with ${\beta}$-divergence, when neither speech nor noise training data is available beforehand. We propose a robust version of non-negative matrix factorization, inspired by the recently developed sparse and low-rank decomposition, in which the data matrix is decomposed into the sum of a low-rank matrix and a sparse matrix. Efficient multiplicative update rules to minimize the ${\beta}$-divergence-based cost function are derived. A convolutional extension of the proposed algorithm is also proposed, which considers the time dependency of the non-negative noise bases. Experimental speech separation results show that the proposed convolutional RNMF successfully separates the repeating time-varying spectral structures from the magnitude spectrum of the mixture, and does so without any prior training.

• Journal of information and communication convergence engineering
• /
• v.1 no.4
• /
• pp.209-212
• /
• 2003

In this paper, we proposed new speech feature parameter through parts-based feature extraction of speech spectrum using Non-Negative Matrix Factorization (NMF). NMF can effectively reduce dimension for multi-dimensional data through matrix factorization under the non-negativity constraints, and dimensionally reduced data should be presented parts-based features of input data. For speech feature extraction, we applied Mel-scaled filter bank outputs to inputs of NMF, than used outputs of NMF for inputs of speech recognizer. From recognition experiment result, we could confirm that proposed feature parameter is superior in recognition performance than mel frequency cepstral coefficient (MFCC) that is used generally.

• Proceedings of the IEEK Conference
• /
• 2003.11a
• /
• pp.49-52
• /
• 2003

In this paper, we propose new speech feature parameter using NMf(Non-Negative Matrix Factorization). NMF can represent multi-dimensional data based on effective dimensional reduction through matrix factorization under the non-negativity constraint, and reduced data present parts-based features of input data. In this paper, we verify about usefulness of NMF algorithm for speech feature extraction applying feature parameter that is got using NMF in Mel-scaled filter bank output. According to recognition experiment result, we could confirm that proposal feature parameter is superior in recognition performance than MFCC(mel frequency cepstral coefficient) that is used generally.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.4
• /
• pp.729-736
• /
• 2019

In this paper, we propose a method to improve the background sound separation performance by adding a Wiener filter to the end of the non - negative matrix factorization method. In the case of a mixed voice signal with background sound, a part that has not yet been completely separated may remain in the signal that separated first by the non-negative matrix factorization method. In this case, it can be reduced in proportion to the size of the residual signal due to the Wiener filter, so that the background sound separation or reduction effect can be expected. Experimental results show that the addition of the Wiener filter is more effective than the case of applying the non-negative matrix factorization method.

• Journal of Ubiquitous Convergence Technology
• /
• v.2 no.1
• /
• pp.18-26
• /
• 2008

The invariance relation existing in the non-negative matrix factorization (NMF) is used for constructing robust image hashes in this work. The image is first re-scaled to a fixed size. Low-pass filtering is performed on the luminance component of the re-sized image to produce a normalized matrix. Entries in the normalized matrix are pseudo-randomly re-arranged under the control of a secret key to generate a secondary image. Non-negative matrix factorization is then performed on the secondary image. As the relation between most pairs of adjacent entries in the NMF's coefficient matrix is basically invariant to ordinary image processing, a coarse quantization scheme is devised to compress the extracted features contained in the coefficient matrix. The obtained binary elements are used to form the image hash after being scrambled based on another key. Similarity between hashes is measured by the Hamming distance. Experimental results show that the proposed scheme is robust against perceptually acceptable modifications to the image such as Gaussian filtering, moderate noise contamination, JPEG compression, re-scaling, and watermark embedding. Hashes of different images have very low collision probability. Tampering to local image areas can be detected by comparing the Hamming distance with a predetermined threshold, indicating the usefulness of the technique in digital forensics.

• The KIPS Transactions:PartB
• /
• v.13B no.6 s.109
• /
• pp.625-632
• /
• 2006

Collaborative filtering is a technology that aims at teaming predictive models of user preferences. Collaborative filtering systems have succeeded in Ecommerce market but they have shortcomings of high dimensionality and sparsity. In this paper we propose the nearest neighbor collaborative filtering method using non-negative matrix factorization(NNMF). We replace the missing values in the user-item matrix by using the user variance coefficient method as preprocessing for matrix decomposition and apply non-negative factorization to the matrix. The positive decomposition method using the non-negative decomposition represents users as semantic vectors and classifies the users into groups based on semantic relations. We compute the similarity between users by using vector similarity and selects the nearest neighbors based on the similarity. We predict the missing values of items that didn't rate by a new user based on the values that the nearest neighbors rated items.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.2
• /
• pp.102-110
• /
• 2010

Music transcription is extracting pitch (the height of a musical note) and rhythm (the length of a musical note) information from audio file and making a music score. In this paper, we decomposed a waveform into frequency and rhythm components using Non-Negative Matrix Factorization (NMF) and Non-Negative Sparse coding (NNSC) which are often used for source separation and data clustering. And using the subharmonic summation method, fundamental frequency is calculated from the decomposed frequency components. Therefore, the accurate pitch of each score can be estimated. The proposed method successfully performed music transcription with its results superior to those of the conventional methods which used either NMF or NNSC.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.25 no.3
• /
• pp.107-116
• /
• 2021

We introduce optimization algorithms using Bregman Divergence for solving non-negative matrix factorization (NMF) problems. Bregman divergence is known a generalization of some divergences such as Frobenius norm and KL divergence and etc. Some algorithms can be applicable to not only NMF with Frobenius norm but also NMF with more general Bregman divergence. Matrix Factorization is a popular non-convex optimization problem, for which alternating minimization schemes are mostly used. We develop the Bregman proximal gradient method applicable for all NMF formulated in any Bregman divergences. In the derivation of NMF algorithm for Bregman divergence, we need to use majorization/minimization(MM) for a proper auxiliary function. We present algorithmic aspects of NMF for Bregman divergence by using MM of auxiliary function.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2004.11a
• /
• pp.117-123
• /
• 2004

마이크로어레이 (microarray) 기술이 개발된 후로 연관된 유전자 클러스터 (cluster)를 찾는 문제는 깊이 연구되어왔다. 이 문제는 핵심적인 과제 중 하나는 생물학적으로 타당한 클러스터의 수를 결정하는 데 있다. 본 논문은 최적의 클러스터 수를 결정하는 기준을 제시하고, non-negative factorization (NMF)를 이용해 클러스터 centroid의 패턴을 찾는 방법을 제안한다. NMF에 의해 발견된 각각의 패턴은 생물학적 프로세스의 특정 부분으로 해석될 수 있다. NMF는 factor matrix의 entity를 non-negative로 제약 (constraint)하고, 이 제약은 오직 additive combination만 허용하기 때문에 이러한 부분적인 패턴을 찾아낼 수 있다. NMF의 유용성은 이미지 분석과 텍스트 분석에서 이미 입증되어 있다. 본 논문에서 제안한 방법에 의해 위의패턴과 유사한 발현 패턴을 갖는 유전자를 모을 수 있었다. 제안된 방법은 human fibroblast데이터와 yeast cell cycle 데이터에 적용해 성능을 입증하였다.