Location:
Search - SVM for speech
Search list
Description: SVMhmm: Learns a hidden Markov model from examples. Training examples (e.g. for part-of-speech tagging) specify the sequence of words along with the correct assignment of tags (i.e. states). The goal is to predict the tag sequences for new sentences.
Platform: |
Size: 95232 |
Author: 王强 |
Hits:
Description: 介绍基于SVM模型的语音识别方法的处理与研究，通过把SVM 与DCT相结合来进行语音的识别。-Introduced the SVM model based on speech recognition method of treatment and research, through the combination of SVM with the DCT for speech recognition.
Platform: |
Size: 626688 |
Author: kellan |
Hits:
Description: 这是一个很好的支持向量机工具箱，它可用于模式识别，图像识别，文字识别，语音识别和手写体识别等领域。-This is a very good support vector machine toolbox, it can be used for pattern recognition, image recognition, text recognition, speech recognition and handwriting recognition and other fields.
Platform: |
Size: 4445184 |
Author: 李全林 |
Hits:
Description: SVMstruct is a Support Vector Machine (SVM) algorithm for predicting multivariate or structured outputs. It performs supervised learning by approximating a mapping
h: X --> Y
using labeled training examples (x1,y1), ..., (xn,yn). Unlike regular SVMs, however, which consider only univariate predictions like in classification and regression, SVMstruct can predict complex objects y like trees, sequences, or sets. Examples of problems with complex outputs are natural language parsing, sequence alignment in protein homology detection, and markov models for part-of-speech tagging. The SVMstruct algorithm can also be used for linear-time training of binary and multi-class SVMs under the linear kernel.
-SVMstruct is a Support Vector Machine (SVM) algorithm for predicting multivariate or structured outputs. It performs supervised learning by approximating a mapping
h: X--> Y
using labeled training examples (x1,y1), ..., (xn,yn). Unlike regular SVMs, however, which consider only univariate predictions like in classification and regression, SVMstruct can predict complex objects y like trees, sequences, or sets. Examples of problems with complex outputs are natural language parsing, sequence alignment in protein homology detection, and markov models for part-of-speech tagging. The SVMstruct algorithm can also be used for linear-time training of binary and multi-class SVMs under the linear kernel.
Platform: |
Size: 109568 |
Author: jon |
Hits:
Description: SVMstruct is a Support Vector Machine (SVM) algorithm for predicting multivariate or structured outputs. It performs supervised learning by approximating a mapping
h: X --> Y
using labeled training examples (x1,y1), ..., (xn,yn). Unlike regular SVMs, however, which consider only univariate predictions like in classification and regression, SVMstruct can predict complex objects y like trees, sequences, or sets. Examples of problems with complex outputs are natural language parsing, sequence alignment in protein homology detection, and markov models for part-of-speech tagging. The SVMstruct algorithm can also be used for linear-time training of binary and multi-class SVMs under the linear kernel.
-SVMstruct is a Support Vector Machine (SVM) algorithm for predicting multivariate or structured outputs. It performs supervised learning by approximating a mapping
h: X--> Y
using labeled training examples (x1,y1), ..., (xn,yn). Unlike regular SVMs, however, which consider only univariate predictions like in classification and regression, SVMstruct can predict complex objects y like trees, sequences, or sets. Examples of problems with complex outputs are natural language parsing, sequence alignment in protein homology detection, and markov models for part-of-speech tagging. The SVMstruct algorithm can also be used for linear-time training of binary and multi-class SVMs under the linear kernel.
Platform: |
Size: 117760 |
Author: jon |
Hits:
Description: 本程序为EMD-HHT-M源代码，供大家交流学习，语音识别专用代码-This program is EMD-HHT-M source code for all to share learning, speech recognition-specific code
Platform: |
Size: 125952 |
Author: wrd |
Hits:
Description: 利用SVM算法来进行中文文本的分类，如一句话里有各种词性的词语则可以进行分类处理-Using SVM algorithm for Chinese text categorization, such as a word in a variety of terms can be part of speech classification
Platform: |
Size: 378880 |
Author: 王寅 |
Hits:
Description: 这是一个GMM模型的语音识别代码 非常详细 而且实用。-This is a GMM model for speech recognition is very detailed and useful code.
Platform: |
Size: 35840 |
Author: zhangge |
Hits:
Description: 建立了普通话语音性别数据库,提出联合梅尔频率频谱系数(Mel2f requency Cep st rum
Coefficient s , MFCC) 的特征提取方法和支持向量机(Support Vector Machine , SVM) 的分类方法进行说话人性别识别,并与其它分类方法进行比较。-A Chinese speech ( mandarin ) database was established for speaker s gender recognition. A
combination met hod is p roposed for gender recognition of speaker s based on support vector machine and
Mel2f requency cep st rum coefficient s (MFCC) for classification and feat ure ext raction respectively.
Platform: |
Size: 303104 |
Author: wangqipeng |
Hits:
Description: Feature Extraction for Speech Processing and SVM Classification of Voice Samples. This package contains Feature Extraction and Classification Matlab codes and some turkish voice records.
Platform: |
Size: 1234944 |
Author: Merdem |
Hits:
Description: 调信息在汉语语音识别中具有非常重要的意义。采用支持向量机对连续汉语连续语音进行声调识别实
验，首先采用基于Teager能量算子和过零率的两级判别策略对连续语音进行浊音段提取，然后建立了适合于支持向
量机分类模型的等维声调特征向量。使用6个二类SVM模型对非特定人汉语普通话的4种声调进行分类识别，与
BP神经网络相比，支持向量杌具有更高的识别率。-Tone is an essential component for word formation in Chinese languages．It plays a very important role in the
transmission of information in speech communication．We looked at using support vector machines(SVMs)for auto—
matic tone recognition in continuously spoken Mandarin．The voiced segments were detected based on Teager Energy
Operation and ZCIL Compared with BP neural network。considerable improvement was achieved by adopting 6 binary-
SVMs scheme in a speaker-independent Mandarin tone recognition system．
Platform: |
Size: 316416 |
Author: 沈 |
Hits:
Description: 针对稀疏表示识别方法需要大量样本训练过完备字典且特征冗余度较高的问题，提出了结合过完备字典学习与PCA降维的小样本语音情感识别算法.该方法首先用PCA降维方法将特征降维，再将处理后的特征用于过完备字典训练与稀疏表示识别方法，从而给出了语音情感特征的稀疏表示方法，并确定了新算法的具体步骤.为验证其有效性，在同等特征维数下，将方法与BP, SVM进行比较，并对比、分析语音情感特征稀疏化前后对语音情感识别率、时间效率以及空间效率的影响.试验结果表明，所提出方法的识别率比SVM与BP高 与采用稀疏化前的特征相比，稀疏化后的特征向量更便于处理，平均识别率提高约15 ，时间效率提高近原来的1 /2，空间效率提升近原来的1 /3.
-Identification methods for sparse representation requires a lot of training samples and high over-complete dictionary feature redundancy problem, a combination of over-complete dictionary learning and PCA dimension small sample speech emotion recognition algorithms. Firstly, the PCA dimension reduction methods feature reduction, feature and then treatment for the over-complete dictionary training and recognition sparse representation, which gives a speech emotion feature sparse representation, and to determine the specific steps of the new algorithm. To verify its validity, in Under the same number of features, the method and BP, SVM compare and contrast, analyze the impact before and after the speech emotion feature sparse speech emotion recognition rate, time-efficient and space-efficient. experimental results show that the recognition rate of the proposed method than High SVM and BP compared to pre-thinning characteristics using eigenvectors easier after thinning processing, the av
Platform: |
Size: 629760 |
Author: wangming |
Hits:
Description: 改进的育群算法，结合SVM用于语音识别,,识别率明显提高-Improved fertility swarm algorithm, combined with SVM for speech recognition
Platform: |
Size: 1024 |
Author: 朱文婧 |
Hits:
Description: vad_directed_by_noise_classification.m
This code is an implementation of VAD algorithm proposed in:
Robust voice activity detection directed by noise classification
please cite the article in your paper:
Robust voice activity detection directed by noise classification, J Saeedi, SM Ahadi, K Faez
Signal, Image and Video Processing, 1-12
the folder is also contained the following
* different noise models for svm
* different sub_functions.
* three speech signal TIMIT dataset and their vad labels
Note that you need to download noise dataset from
http://spib.rice.edu/spib/select.
and libsvm toolbox from
http://www.csie.ntu.edu.tw/~cjlin/libsvm/
It should be mentioned that both speech and noise should be sampled at 8 KHz.
Jamal Saeedi
Amirkabir University of Technology
Electrical Engineering Department-vad_directed_by_noise_classification.m
This code is an implementation of VAD algorithm proposed in:
Robust voice activity detection directed by noise classification
please cite the article in your paper:
Robust voice activity detection directed by noise classification, J Saeedi, SM Ahadi, K Faez
Signal, Image and Video Processing, 1-12
the folder is also contained the following
* different noise models for svm
* different sub_functions.
* three speech signal TIMIT dataset and their vad labels
Note that you need to download noise dataset from
http://spib.rice.edu/spib/select.
and libsvm toolbox from
http://www.csie.ntu.edu.tw/~cjlin/libsvm/
It should be mentioned that both speech and noise should be sampled at 8 KHz.
Jamal Saeedi
Amirkabir University of Technology
Electrical Engineering Department
Platform: |
Size: 1024 |
Author: Ilya |
Hits:
Description: In this method voice activity detection (VAD) is formulated as a two class classification problem using support vector machines (SVM). The proposed method combines a noise robust feature extraction process together with SVM models trained in different background noises for speech/nonspeech classification. A multi-class SVM is also used to classify background noises in order to SVM model for VAD algorithm. The proposed VAD is tested with TIMIT data artificially distorted by different additive noise types.
Platform: |
Size: 10744832 |
Author: jamal |
Hits:
Description: Complete HMM-based speech recognition system, For feature extraction, signal de-noising, Analysis of the signal time domain, frequency domain, cepstrum, cyclic spectrum, etc.
Platform: |
Size: 59392 |
Author: 徐在
|
Hits: