Spoken Digits Recognition using Weighted MFCC and Improved Features for Dynamic Time Warping

Chapaneri, Santosh

doi:10.5120/5022-7167

Cited by 34 publications

(16 citation statements)

References 32 publications

(39 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…RS Sound Modality: Similar to [33,34], Mel-frequency cepstral coefficients (MFCC) are utilized to delegate the RS voices, because MFCC uses cepstrum feature extraction [35,36], which is more in line with the principle of human hearing. So it is the most common and effective voice feature extraction algorithm [37,38].…”

Section: Rs Image Modalitymentioning

confidence: 99%

A Deep Hashing Technique for Remote Sensing Image-Sound Retrieval

Chen

2019

Remote Sensing

View full text Add to dashboard Cite

With the rapid progress of remote sensing (RS) observation technologies, cross-modal RS image-sound retrieval has attracted some attention in recent years. However, these methods perform cross-modal image-sound retrieval by leveraging high-dimensional real-valued features, which can require more storage than low-dimensional binary features (i.e., hash codes). Moreover, these methods cannot directly encode relative semantic similarity relationships. To tackle these issues, we propose a new, deep, cross-modal RS image-sound hashing approach, called deep triplet-based hashing (DTBH), to integrate hash code learning and relative semantic similarity relationship learning into an end-to-end network. Specially, the proposed DTBH method designs a triplet selection strategy to select effective triplets. Moreover, in order to encode relative semantic similarity relationships, we propose the objective function, which makes sure that that the anchor images are more similar to the positive sounds than the negative sounds. In addition, a triplet regularized loss term leverages approximate l1-norm of hash-like codes and hash codes and can effectively reduce the information loss between hash-like codes and hash codes. Extensive experimental results showed that the DTBH method could achieve a superior performance to other state-of-the-art cross-modal image-sound retrieval methods. For a sound query RS image task, the proposed approach achieved a mean average precision (mAP) of up to 60.13% on the UCM dataset, 87.49% on the Sydney dataset, and 22.72% on the RSICD dataset. For RS image query sound task, the proposed approach achieved a mAP of 64.27% on the UCM dataset, 92.45% on the Sydney dataset, and 23.46% on the RSICD dataset. Future work will focus on how to consider the balance property of hash codes to improve image-sound retrieval performance.

show abstract

Section: Rs Image Modalitymentioning

confidence: 99%

A Deep Hashing Technique for Remote Sensing Image-Sound Retrieval

Chen

2019

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…System is found successful and can identify spoken digit at 89.2% recognition rate, which is well acceptable rate of accuracy for speech recognition. In [10], a novel technique was proposed for feature parameter extraction based on MFCC and feature recognition using dynamic time warping algorithm for application in speaker-independent isolated digits recognition. Using the proposed Weighted MFCC (WMFCC), they compute the local and global features using Improved Features for DTW algorithm (IFDTW), rather than using the pure feature values or their estimated derivatives.…”

Section: Related Workmentioning

confidence: 99%

Amazigh Isolated-Word speech recognition system using Hidden Markov Model toolkit (HTK)

Elouahabi¹,

Atounti²,

Bellouki³

2016

2016 International Conference on Information Technology for Organizations Development (IT4OD)

View full text Add to dashboard Cite

This paper aims to build a speaker-independent automatic Amazigh Isolated-Word speech recognition system. Hidden Markov Model toolkit (HTK) that uses hidden Markov Models has been used to develop the system. The recognition vocabulary consists on the Amazigh Letters and Digits. The system has been trained to recognize the Amazigh 10 first digits and 33 alphabets. Mel frequency spectral coefficients (MFCCs) have been used to extract the feature. The training data has been collected from 60 speakers including both males and females. The test-data used for evaluating the system-performance has been collected from 20 speakers. The experimental results show that the presented system provides the overall word-accuracy 80%. The initial results obtained are very satisfactory in comparison with the training database's size, this encourages us to increase system performance to achieve a higher recognition rate. Keywords-automatic speech recognition;hidden markov models; mel frequency spectral coefficients;hidden markov model toolkit (HTK); amazigh language.

show abstract

“…concentration of more energy in small number of coefficients as compared to other transform, (2) less correlation among the generated MFCC values as a result of DCT in which the lower order MFCC values are observed to represent the smooth spectral shape while the higher order MFCC values are observed to represent the excitation signal etc. [4]. The resulting MFCC coefficients are given by Eq.…”

Section: Fig 1: Mfcc Flowchartmentioning

confidence: 99%

“…The arrangement of the Mel-filter bank is such that the edges of every band coincide with the center of the corresponding neighboring band which can be observed in Fig. 2 [4]. An approximation to the Mel-filter bank is a bank of linearly spaced filters with equal bandwidth under 1000 Hz and logarithmic spacing above 1000 Hz such that the center frequency of each filter is 1.1 times the preceding center frequency [16].…”

Section: Fig 1: Mfcc Flowchartmentioning

confidence: 99%

“…Features that are typically extracted from speech can be broadly classified into vocal tract spectrum features, prosodic features and nonlinear features among the others. Some of the distinct vocal tract spectrum features (also called as segmental features) to be extracted in the research of SER includes Mel Frequency Cepstral Coefficients (MFCC) [3,4], Linear Prediction Cepstral Coefficients (LPCC) [5], Log Frequency Power Coefficients (LFPC), Mel Energy Spectrum Dynamic Coefficients (MEDC), Delta-Spectral Cepstral Coefficients (DSCC) [6] etc. Prosodic features constitute pitch, jitter, energy, spectral tilt, duration etc.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Emotion Recognition from Speech using Discriminative Features

Chandrasekar¹,

Chapaneri²,

Jayaswal³

2014

IJCA

View full text Add to dashboard Cite

Creating an accurate Speech Emotion Recognition (SER) system depends on extracting features relevant to that of emotions from speech. In this paper, the features that are extracted from the speech samples include Mel Frequency Cepstral Coefficients (MFCC), energy, pitch, spectral flux, spectral roll-off and spectral stationarity. In order to avoid the 'curse of dimensionality', statistical parameters, i.e. mean, variance, median, maximum, minimum, and index of dispersion have been applied on the extracted features. For classifying the emotion in an unknown test sample, Support Vector Machines (SVM) has been chosen due to its proven efficiency. Through experimentation on the chosen features, an average classification accuracy of 86.6% has been achieved using one-v/s-all multi-class SVM which is further improved to 100% when reduced to binary form problem. Classifier metrics viz. precision, recall, and F-score values show that the proposed system gives improved accuracy for Emo-DB. General TermsPattern Recognition

show abstract

Spoken Digits Recognition using Weighted MFCC and Improved Features for Dynamic Time Warping

Cited by 34 publications

References 32 publications

A Deep Hashing Technique for Remote Sensing Image-Sound Retrieval

A Deep Hashing Technique for Remote Sensing Image-Sound Retrieval

Amazigh Isolated-Word speech recognition system using Hidden Markov Model toolkit (HTK)

Emotion Recognition from Speech using Discriminative Features

Contact Info

Product

Resources

About