22 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 10 / 22
Yayın A New speech coding algorithm using zero cross and phoneme based SYMPES(IEEE, 2013-07-11) Şişman, Burak; Güz, Ümit; Gürkan, Hakan; Yarman, Bekir Sıddık BinboğaIn this work, a new low bit rate hybrid speech coding approach which combines the benefits of the SYMPES (Systematic Procedure for Predefined Envelope and Signature Sequences) and zero cross and phoneme based segmentation is proposed. In the new approach, the SYMPES structure is developed in the phoneme based fashion. In order to achieve lower bit rates, some drawbacks such as computational complexity, relatively high encoding times etc. of the SYMPES are also eliminated in the new version. Experimental results show that in almost same bit rates very promising speech quality is obtained compared to the other conventional methods such as CELP (Code Excited Linear Predictive) coding algorithm.Yayın A novel computed tomography image compression method based on classified energy and pattern blocks(IEEE, 2013) Gökbay, İnci Zaim; Gezer, Murat; Güz, Ümit; Gürkan, Hakan; Yarman, Bekir Sıddık BinboğaIn this work, a new biomedical image compression method is proposed based on the classified energy and pattern blocks (CEPB). CEPB based compression method is specifically applied on the Computed Tomography (CT) images and the evaluation results are presented. Essentially, the CEPB is uniquely designed and structured codebook which is located on the both the transmitter and receiver part of a communication system in order to implement encoding and decoding processes. The encoding parameters are block scaling coefficient (BSC) and the index numbers of energy (IE) and pattern blocks (IP) determined for each block of the input images based on the CEPB. The evaluation results show that the newly proposed method provides considerable image compression ratios and image quality.Yayın Model adaptation for dialog act tagging(IEEE, 2006) Tür, Gökhan; Güz, Ümit; Hakkani Tür, DilekIn this paper, we analyze the effect of model adaptation for dialog act tagging. The goal of adaptation is to improve the performance of the tagger using out-of-domain data or models. Dialog act tagging aims to provide a basis for further discourse analysis and understanding in conversational speech. In this study we used the ICSI meeting corpus with high-level meeting recognition dialog act (MRDA) tags, that is, question, statement, backchannel, disruptions, and floor grabbers/holders. We performed controlled adaptation experiments using the Switchboard (SWBD) corpus with SWBD-DAMSL tags as the out-of-domain corpus. Our results indicate that we can achieve significantly better dialog act tagging by automatically selecting a subset of the Switchboard corpus and combining the confidences obtained by both in-domain and out-of-domain models via logistic regression, especially when the in-domain data is limited.Yayın Extension of conventional co-training learning strategies to three-view and committee-based learning strategies for effective automatic sentence segmentation(IEEE, 2018) Dalva, Doğan; Güz, Ümit; Gürkan, HakanThe objective of this work is to develop effective multi-view semi-supervised machine learning strategies for sentence boundary classification problem when only small sets of sentence boundary labeled data are available. We propose three-view and committee-based learning strategies incorporating with co-training algorithms with agreement, disagreement, and self-combined learning strategies using prosodic, lexical and morphological information. We compare experimental results of proposed three-view and committee-based learning strategies to other semi-supervised learning strategies in the literature namely, self-training and co-training with agreement, disagreement, and self-combined strategies. The experiment results show that sentence segmentation performance can be highly improved using multi-view learning strategies that we propose since data sets can be represented by three redundantly sufficient and disjoint feature sets. We show that the proposed strategies substantially improve the average performance when only a small set of manually labeled data is available for Turkish and English spoken languages, respectively.Yayın A new speech modeling method: SYMPES(IEEE, 2006) Güz, Ümit; Gürkan, Hakan; Yarman, Bekir Sıddık BinboğaIn this paper, the new method of speech modeling which is called SYMPES is introduced and it is compared with the commercially available methods. It is shown that for the same compression ratio or better, SYMPES yields considerably better hearing quality over the coders such as G.726 at 16 Kbps and voice excited LPC-10E of 2.4Kbps.Yayın An efficient ECG data compression technique based on predefined signature and envelope vector banks(IEEE, 2005) Gürkan, Hakan; Güz, Ümit; Yarman, Bekir Sıddık BinboğaIn this paper, a new method to compress ElectroCardioGram (ECG) Signal by means of "Predefined Signature and Envelope Vector Banks-PSEVB" is presented. In this work, on a frame basis, any ECG signal is modeled by multiplying three parameters as called the Predefined Signature Vector, Predefined Envelope Vector, and Frame-Scaling Coefficient. It has been demonstrated that the predefined signature vectors and predefined envelope vectors constitute a "PSEVB" to describe any measured ECG signal. In this case, ECG signal for each frame is described in terms of the two indices "R" and "K" of PSEVB and the frame-scaling coefficient. The new compression method achieve good compression ratios with low level reconstruction error while preserving diagnostic information in the reconstructed ECG signal. Furthermore, once PSEVB are stored on each communication node, transmission of ECG signals reduces to the transmission of indexes "R" and "K" of PSEVB and the frame-scaling coefficient which also result in considerable saving in the transmission band.Yayın A new algorithm for high speed speech and audio coding(IEEE, 2007) Güz, Ümit; Gürkan, Hakan; Yarman, Bekir Sıddık BinboğaIn this work, a new mathematical modeling approach is proposed for the representation of the speech and audio signals. This approach is based on the generation of the so called Predefined Signature Sequence (PSS) and Predefined Envelope Sequence (PES) Sets. After the generation process of the PSS and PES sets, they are clustered by effective k-means clustering algorithm and the PSS and PES are redefined by using the centroids of the clusters. By using this approach, the drawbacks such as the size of the sets, speed of the reconstruction process (computational complexity) which arise in our proposed methods previously are highly eliminated. In spite of these improvements, the initial results proved that, the quality of the reconstructed signals remains within the limitations of the acceptable hearing quality.Yayın A new coding method for speech and audio signals(IEEE, 2005) Güz, Ümit; Gürkan, Hakan; Yarman, Bekir Sıddık BinboğaIn this paper a new representation or modeling method of speech signals is introduced. The proposed method is based on the generation of the so-called Predefined Signature S={S R } and Envelope vector E={E K } Sets (PSEVS). These vector sets are speaker and language independent. In this method, once the speech signals are divided into frames with selected lengths, then each frame signal piece X i is reconstructed by means of the mathematical form of X i =C i E K S R . In this representation, C i is called the frame coefficient, S R and E K are the vectors properly assigned from the PSEVS respectively. It is shown that the proposed method provides fast reconstruction and substantial compression ratio with acceptable hearing quality.Yayın A novel method to represent ECG signals via predefined personalized signature and envelope functions(IEEE, 2001) Yarman, Bekir Sıddık Binboğa; Gürkan, Hakan; Güz, Ümit; Aygün, B.In this paper, a new method to model ECG signals by means of "Predefined Personalized Signature and Envelope Functions" is presented. ECG signals are somewhat unique to a person. Moreover, it presents quasi-stationary behavior. Therefore in this work, on a frame basis, personal ECG signals X-i(t) is modeled by the form of X-i(t) approximate to C(i)phi(i)(t) alpha(i)(t). In this model, phi(i)(t) is defined as the Personalized Signature Function (PSF); alpha(i)(t) is referred to as Personalized Envelope Function (PEF) and C-i is called the Frame-Scaling Coefficient (FSC). It has been demonstrated that for each person, the sets Phi = {phi(k)(t)} and A = {alpha(r)(t)} constitute a "Predefined Personalized Functional Bases or Banks (PPFB)" to describe any measured ECG signal. Almost optimum forms of (PPFB), namely {alpha(r)(t)}, {phi(k)(t)} pairs are generated in the Least Mean Square (LMS) sense. Thus, ECG signal for each frame is described in terms of the two indices "R" and "K" of PPFB and the frame-scaling coefficient Ci. It has been shown that the new method of modeling provides significant data compression. Furthermore, once PPFB are stored on each communication node, transmission of ECG signals reduces to the transmission of indexes "R" and "K" of pairs and the coefficients C-i, which also result in considerable saving in the transmission band.Yayın A novel noise robust and low bit rate speech coding algorithm(IEEE, 2009) Güz, Ümit; Gürkan, Hakan; Yarman, Bekir Sıddık BinboğaIn this work, a new noise robust and variable length frame based speech modeling method is introduced. This method consists of three major steps which includes noise removal algorithm, coding and encoding algorithms, respectively. Coding and encoding parts are developed based on SYMPES (SYsteMatic Procedure for Predefined Envelope and Signature sequence sets). These sets have been developed in two types which represent voiced and unvoiced parts of the speech signals separately in order to obtain more efficient coding strategy and higher compression ratio while preserving the perceptual quality of the speech signals. As an extension of our previous works our new framework is not only consider the coding of the clean speech signals but also noisy speech signals. The new noise robust module suppresses the noise and delivers the clean speech signal to the newly designed modeling part. The modeling part promises higher compression ratios by switching to the more appropriate type of predefined sets take into account the voiced and unvoiced frames.
- «
- 1 (current)
- 2
- 3
- »
