Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6488011
S. Shahnawazuddin, Deepak Thotappa, B. Sarma, A. Deka, S. Prasanna, R. Sinha
In this work, a spoken query system developed for accessing the price of agricultural commodities in Assamese language is described. The developed system enables the user to access the latest price of the commodity by calling the system using a landline/mobile phone. The spoken query system consists of interactive voice response (IVR) and automatic speech recognition (ASR) modules which are developed using open source resources. For the development of ASR models, the task specific speech data is collected from different dialect regions of Assam. The issues in test data adaptation are highlighted and a constrained data-unseen speaker adaptation approach is implemented which is found to give a relative improvement by 8% in baseline performance.
{"title":"Assamese spoken query system to access the price of agricultural commodities","authors":"S. Shahnawazuddin, Deepak Thotappa, B. Sarma, A. Deka, S. Prasanna, R. Sinha","doi":"10.1109/NCC.2013.6488011","DOIUrl":"https://doi.org/10.1109/NCC.2013.6488011","url":null,"abstract":"In this work, a spoken query system developed for accessing the price of agricultural commodities in Assamese language is described. The developed system enables the user to access the latest price of the commodity by calling the system using a landline/mobile phone. The spoken query system consists of interactive voice response (IVR) and automatic speech recognition (ASR) modules which are developed using open source resources. For the development of ASR models, the task specific speech data is collected from different dialect regions of Assam. The issues in test data adaptation are highlighted and a constrained data-unseen speaker adaptation approach is implemented which is found to give a relative improvement by 8% in baseline performance.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116890043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487990
Harish Padaki, K. Nathwani, R. Hegde
Clean speech acquisition from distant microphones is often affected by the phenomenon of reverberation. In this paper, the significance of a blind single channel dereverberation method using the linear prediction (LP) residual of the reverberated speech signal is proposed. A relation between the LP residual of the clean and the reverberated speech signal is also derived using the acoustic room impulse response. However in the proposed method, the clean LP residual is computed from its reverberated counterpart by a method of cepstral subtraction. Hence there is no estimation of the acoustic room impulse response (AIR) making the method computationally simple. Experiments on speech dereverberation and distant speech recognition are conducted at various direct to reverberant ratios (DRR). The results are presented using objective measures, subjective measures, and word error rates (WER) and compared to methods available in literature to illustrate the significance of this method.
{"title":"Single channel speech dereverberation using the LP residual cepstrum","authors":"Harish Padaki, K. Nathwani, R. Hegde","doi":"10.1109/NCC.2013.6487990","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487990","url":null,"abstract":"Clean speech acquisition from distant microphones is often affected by the phenomenon of reverberation. In this paper, the significance of a blind single channel dereverberation method using the linear prediction (LP) residual of the reverberated speech signal is proposed. A relation between the LP residual of the clean and the reverberated speech signal is also derived using the acoustic room impulse response. However in the proposed method, the clean LP residual is computed from its reverberated counterpart by a method of cepstral subtraction. Hence there is no estimation of the acoustic room impulse response (AIR) making the method computationally simple. Experiments on speech dereverberation and distant speech recognition are conducted at various direct to reverberant ratios (DRR). The results are presented using objective measures, subjective measures, and word error rates (WER) and compared to methods available in literature to illustrate the significance of this method.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125084250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487967
S. Agarwal, N. Pathak, Dharmendra Singh
The paper presents a novel hybrid configuration to obtain dual band operation at millimeter wave frequency in W band (75 GHz to 110 GHz). Dual band operation is achieved by introducing a slot in the main patch, which is coupled to the parasitic patches along non-radiating edges through gap coupling. The antenna resonates at two frequencies, viz., 85 GHz and 94 GHz with fractional bandwidths of 8% and 1.8%, respectively. The simulated gains at two frequencies are 8.28 dB and 8.89 dB, respectively. A scaled prototype of the proposed dual band millimeter wave antenna has been fabricated. Comparison of measured and simulated results has been done in terms of return loss and far field measurements.
{"title":"Concurrent 85GHz/94GHz slotted gap coupled parasitic microstrip antenna for millimeter wave applications","authors":"S. Agarwal, N. Pathak, Dharmendra Singh","doi":"10.1109/NCC.2013.6487967","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487967","url":null,"abstract":"The paper presents a novel hybrid configuration to obtain dual band operation at millimeter wave frequency in W band (75 GHz to 110 GHz). Dual band operation is achieved by introducing a slot in the main patch, which is coupled to the parasitic patches along non-radiating edges through gap coupling. The antenna resonates at two frequencies, viz., 85 GHz and 94 GHz with fractional bandwidths of 8% and 1.8%, respectively. The simulated gains at two frequencies are 8.28 dB and 8.89 dB, respectively. A scaled prototype of the proposed dual band millimeter wave antenna has been fabricated. Comparison of measured and simulated results has been done in terms of return loss and far field measurements.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128183414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6488038
R. Pandya, V. Chandra, D. Chadha
In the present generation optical networks, power economy and impairment awareness for the increased survivability become crucial. On the other hand, there is a trade - off exists between power economy and impairment awareness. To achieve power economy, transparency is introduced in optical networks. Moreover, optical transparency lacks in regeneration of optical signal. This results in an accumulation of noise signal due to the linear and nonlinear impairments of optical media. Furthermore, to overcome the effects of impairments, transmitted power is increased which increases the power cost of the network. Hence, the present paper gives a novel solution which incorporates impairments awareness in the network by enhancing the control plane of the network without increasing the power cost and complexity. This is achieved by Impairment Aware Routing and Wavelength Assignment (IA - RWA) to selective switching off of impaired light paths, which further increases the power economy. We use sleep mode strategy to achieve minimal power cost with SBSPP and saved power is further used in a contention resolution strategy to enhance the data rate.
{"title":"Power impairment aware methodology with segment based shared path protection in survivable optical packet switched networks","authors":"R. Pandya, V. Chandra, D. Chadha","doi":"10.1109/NCC.2013.6488038","DOIUrl":"https://doi.org/10.1109/NCC.2013.6488038","url":null,"abstract":"In the present generation optical networks, power economy and impairment awareness for the increased survivability become crucial. On the other hand, there is a trade - off exists between power economy and impairment awareness. To achieve power economy, transparency is introduced in optical networks. Moreover, optical transparency lacks in regeneration of optical signal. This results in an accumulation of noise signal due to the linear and nonlinear impairments of optical media. Furthermore, to overcome the effects of impairments, transmitted power is increased which increases the power cost of the network. Hence, the present paper gives a novel solution which incorporates impairments awareness in the network by enhancing the control plane of the network without increasing the power cost and complexity. This is achieved by Impairment Aware Routing and Wavelength Assignment (IA - RWA) to selective switching off of impaired light paths, which further increases the power economy. We use sleep mode strategy to achieve minimal power cost with SBSPP and saved power is further used in a contention resolution strategy to enhance the data rate.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130786232","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6488044
Mallikarjun B. Channappagoudar, P. Venkataram
Mobile ad-hoc network is a wireless ad-hoc network with dynamic network topology. The Dynamicity, due to the random node movement, and scarcity of resources lead to a challenge in monitoring the nodes in a MANET. Monitoring the lack of resources (bandwidth, buffer, and energy), misbehavior, and mobility at node level remains, a challenge. In a MANET the proposed protocol uses both static as well as mobile agents, where the mobile agents migrate to different clusters of the zones respectively, collect the node status information periodically, and provide a high level information to the static agent (which resides at the central node) by analyzing the raw information at the nodes. This, in turn, reduces the network traffic and conserves the workload of the central node, where a static agent is available with high level information and in coordination with other modules. The protocol has been tested in different size MANETs with variable number of nodes and applications. The results shown in the simulation indicates the effectiveness of the protocol.
{"title":"Mobile agent based node monitoring protocol for MANETs","authors":"Mallikarjun B. Channappagoudar, P. Venkataram","doi":"10.1109/NCC.2013.6488044","DOIUrl":"https://doi.org/10.1109/NCC.2013.6488044","url":null,"abstract":"Mobile ad-hoc network is a wireless ad-hoc network with dynamic network topology. The Dynamicity, due to the random node movement, and scarcity of resources lead to a challenge in monitoring the nodes in a MANET. Monitoring the lack of resources (bandwidth, buffer, and energy), misbehavior, and mobility at node level remains, a challenge. In a MANET the proposed protocol uses both static as well as mobile agents, where the mobile agents migrate to different clusters of the zones respectively, collect the node status information periodically, and provide a high level information to the static agent (which resides at the central node) by analyzing the raw information at the nodes. This, in turn, reduces the network traffic and conserves the workload of the central node, where a static agent is available with high level information and in coordination with other modules. The protocol has been tested in different size MANETs with variable number of nodes and applications. The results shown in the simulation indicates the effectiveness of the protocol.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131316440","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487980
R. Das, B. K. Das, M. Chakraborty
Combination of two adaptive filters working in parallel for achieving better performance both in term of convergence speed and excess mean square error (EMSE) has been considered by several researchers in recent past. Prominent among these include convex combination (where combinational weight factors are within the range [0 1], while summing up to one), affine combination (where the combinational weight factors are free from any range constraint, while still summing up to one) and unconstrained model combination (where the output of constituent filters are combined using another adaptive algorithm). In this paper, we propose a novel way of using two adaptive filters for achieving better performance, using the cooperative learning approach. For this, we employ one LMS based adaptive filter that uses a larger step size and thus has a faster rate of convergence at the expense of higher EMSE. The other filter employed uses a modified version of the LMS algorithm, which employs a much lesser step size, but has one extra update term in the weight update relation that helps in learning from the faster filter its filter weight information. The learning takes place during the transient phase, while, in the steady state, two filters become almost independent of each other. Presence of the learning component in the weight update recursion enables the filter to converge much faster while a smaller step size ensures much less steady state EMSE. The claims are supported by theoretical as well as detailed simulation studies.
{"title":"Improving the performance of the LMS algorithm via cooperative learning","authors":"R. Das, B. K. Das, M. Chakraborty","doi":"10.1109/NCC.2013.6487980","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487980","url":null,"abstract":"Combination of two adaptive filters working in parallel for achieving better performance both in term of convergence speed and excess mean square error (EMSE) has been considered by several researchers in recent past. Prominent among these include convex combination (where combinational weight factors are within the range [0 1], while summing up to one), affine combination (where the combinational weight factors are free from any range constraint, while still summing up to one) and unconstrained model combination (where the output of constituent filters are combined using another adaptive algorithm). In this paper, we propose a novel way of using two adaptive filters for achieving better performance, using the cooperative learning approach. For this, we employ one LMS based adaptive filter that uses a larger step size and thus has a faster rate of convergence at the expense of higher EMSE. The other filter employed uses a modified version of the LMS algorithm, which employs a much lesser step size, but has one extra update term in the weight update relation that helps in learning from the faster filter its filter weight information. The learning takes place during the transient phase, while, in the steady state, two filters become almost independent of each other. Presence of the learning component in the weight update recursion enables the filter to converge much faster while a smaller step size ensures much less steady state EMSE. The claims are supported by theoretical as well as detailed simulation studies.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131693197","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487965
S. Ashok Kumar, T. Shanmuganantham
In this paper, the CPW fed circular slot antennas are proposed for Industrial, Scientific, and Medical (ISM) band (2.4–2.48 GHz) applications. To make the designed antenna suitable for implantation, it is embedded in biocompatible Al2O3 ceramic substrate. The antenna was simulated by immersing it in a phantom liquid, imitating the electrical properties of the human muscle tissue. A study of the sensitivity of the antenna performance as a function of its dielectric parameters of the environment in which it is immersed was performed. Simulations in various dimensions state demonstrate that the antenna covers the complete ISM band. The demonstration among the design EM characteristics of the antenna is presented by current distributions.
{"title":"Implantable CPW fed circular slot antenna for ISM band","authors":"S. Ashok Kumar, T. Shanmuganantham","doi":"10.1109/NCC.2013.6487965","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487965","url":null,"abstract":"In this paper, the CPW fed circular slot antennas are proposed for Industrial, Scientific, and Medical (ISM) band (2.4–2.48 GHz) applications. To make the designed antenna suitable for implantation, it is embedded in biocompatible Al2O3 ceramic substrate. The antenna was simulated by immersing it in a phantom liquid, imitating the electrical properties of the human muscle tissue. A study of the sensitivity of the antenna performance as a function of its dielectric parameters of the environment in which it is immersed was performed. Simulations in various dimensions state demonstrate that the antenna covers the complete ISM band. The demonstration among the design EM characteristics of the antenna is presented by current distributions.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125567882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487972
A. Deshmukh, Ankita R. Jain, K. Ray
The broadband microstrip antenna is more commonly realized by cutting the slot at an appropriate position inside the patch. The slot is said to introduce a mode near the resonance frequency of the fundamental mode of the patch and yields broadband response. The compact microstrip antenna is realized by placing the shorting post or the plate along the zero field line at the fundamental mode of the patch. In this paper, the compact variation of equilateral triangular microstrip antenna, a shorted sectoral microstrip antenna is discussed. A broadband proximity fed rectangular slot cut shorted sectoral microstrip antenna is proposed. The analysis to study the effects of slot on the broadband response is presented. The slot reduces the higher order mode resonance frequencies of the shorted sectoral patch and along with the fundamental patch mode yields broadband response. The simulated and measured bandwidth of more than 550 MHz is realized. Since the shorted patch is used, the antenna shows radiation pattern in end-fire direction with a gain of more than 6 dBi over the bandwidth.
{"title":"Broadband sectoral slot cut microstrip antenna","authors":"A. Deshmukh, Ankita R. Jain, K. Ray","doi":"10.1109/NCC.2013.6487972","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487972","url":null,"abstract":"The broadband microstrip antenna is more commonly realized by cutting the slot at an appropriate position inside the patch. The slot is said to introduce a mode near the resonance frequency of the fundamental mode of the patch and yields broadband response. The compact microstrip antenna is realized by placing the shorting post or the plate along the zero field line at the fundamental mode of the patch. In this paper, the compact variation of equilateral triangular microstrip antenna, a shorted sectoral microstrip antenna is discussed. A broadband proximity fed rectangular slot cut shorted sectoral microstrip antenna is proposed. The analysis to study the effects of slot on the broadband response is presented. The slot reduces the higher order mode resonance frequencies of the shorted sectoral patch and along with the fundamental patch mode yields broadband response. The simulated and measured bandwidth of more than 550 MHz is realized. Since the shorted patch is used, the antenna shows radiation pattern in end-fire direction with a gain of more than 6 dBi over the bandwidth.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126586894","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487924
Sachin Kadam, G. Sharma, R. Bansal
Spectrum sensing is an important problem in cognitive radio. Outliers present in the channel deteriorate the performance of existing non-robust algorithms. We consider the problem of limiting the influence of outliers in cooperative spectrum sensing techniques. In this work we use Huber's least favorable pair based on mixture model with appropriate nominal distributions in the DualCUSUM algorithm, a sequential change point detection algorithm used for spectrum sensing. We show that proposed robust DualCUSUM algorithm performs better than existing DualCUSUM algorithm in the presence of outliers. It is also shown by simulation results that better performance can be achieved when the design parameter used in obtaining the least favorable pair equals the actual contamination level in the data. A method to generate random numbers which follow least favorable pair of distributions is also discussed.
{"title":"Robust dual cumulative sum algorithm for cooperative spectrum sensing","authors":"Sachin Kadam, G. Sharma, R. Bansal","doi":"10.1109/NCC.2013.6487924","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487924","url":null,"abstract":"Spectrum sensing is an important problem in cognitive radio. Outliers present in the channel deteriorate the performance of existing non-robust algorithms. We consider the problem of limiting the influence of outliers in cooperative spectrum sensing techniques. In this work we use Huber's least favorable pair based on mixture model with appropriate nominal distributions in the DualCUSUM algorithm, a sequential change point detection algorithm used for spectrum sensing. We show that proposed robust DualCUSUM algorithm performs better than existing DualCUSUM algorithm in the presence of outliers. It is also shown by simulation results that better performance can be achieved when the design parameter used in obtaining the least favorable pair equals the actual contamination level in the data. A method to generate random numbers which follow least favorable pair of distributions is also discussed.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115859982","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-28DOI: 10.1109/NCC.2013.6487994
cBhargav Srinivas Ch, cNeethu Mariam Joy, cRaghavendra R. Bilgi, C. Umesh
In this paper we propose an adaptive training method for parameter estimation of acoustic models in the speech recognition system. Our technique is inspired from the Cluster Adaptive Training (CAT) method which is used for rapid speaker adaptation. Instead of adapting the model to a speaker as in CAT, we adapt the parameters of the context dependent triphone states (tied states) from context independent states (monophones). This is achieved by finding a global mapping of parameters of the tied state from the parametric subspace of monophone models. This technique is similar to Subspace Gaussian Mixture Model (SGMM), but differs in the initialization of parameters and in the update of weights of Gaussian mixture components. We show that, the proposed method can match the performance of the conventional HMM system for large amount of training data and outperforms it when the number of training examples are less.
{"title":"Subspace modeling technique using monophones for speech recognition","authors":"cBhargav Srinivas Ch, cNeethu Mariam Joy, cRaghavendra R. Bilgi, C. Umesh","doi":"10.1109/NCC.2013.6487994","DOIUrl":"https://doi.org/10.1109/NCC.2013.6487994","url":null,"abstract":"In this paper we propose an adaptive training method for parameter estimation of acoustic models in the speech recognition system. Our technique is inspired from the Cluster Adaptive Training (CAT) method which is used for rapid speaker adaptation. Instead of adapting the model to a speaker as in CAT, we adapt the parameters of the context dependent triphone states (tied states) from context independent states (monophones). This is achieved by finding a global mapping of parameters of the tied state from the parametric subspace of monophone models. This technique is similar to Subspace Gaussian Mixture Model (SGMM), but differs in the initialization of parameters and in the update of weights of Gaussian mixture components. We show that, the proposed method can match the performance of the conventional HMM system for large amount of training data and outperforms it when the number of training examples are less.","PeriodicalId":202526,"journal":{"name":"2013 National Conference on Communications (NCC)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121034752","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：