Feature Extraction Based on Wavelet Domain Hidden Markov Tree Model for Robust Speech Recognition

Jung, Sungyun; Son, Jong‐Yeong; Bae, Kyung‐Yeol

doi:10.1007/978-3-540-30549-1_116

Cited by 3 publications

(3 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, the HMT model is also suitable for robust ASR. In fact, an enhanced method for the utilisation of the HMT model for de‐noising applications is presented in [76]. The proposed method is made up of two cascaded stages: the first one being the HMT model for the de‐noising process and at the second stage a weighted filter bank analysis is performed.…”

Section: Automatic Speech Recognition Systemsmentioning

confidence: 99%

Comparative study of automatic speech recognition techniques

Cutajar

Gatt

Grech

et al. 2013

IET signal process.

View full text Add to dashboard Cite

Over the past decades, extensive research has been carried out on various possible implementations of automatic speech recognition (ASR) systems. The most renowned algorithms in the field of ASR are the mel-frequency cepstral coefficients and the hidden Markov models. However, there are also other methods, such as wavelet-based transforms, artificial neural networks and support vector machines, which are becoming more popular. This review article presents a comparative study on different approaches that were proposed for the task of ASR, and which are widely used nowadays. † training time increases linearly with increase in vocabulary size [42] † quantisation error in the discrete representation of speech signals [42] † temporal information is ignored [42] PCA † reduction in the feature vector's size, while retaining much of the significant information [131] † robust [59, 60] † computationally expensive for high-dimensional data [8] LDA † maximises the distance between classes, but minimises the within class distance [132] † robust [133] † sample distribution is assumed a priori to be Gaussian [63] † class samples are assumed to have equal variance [63] Classification technique Advantages Disadvantages HMM † able to model time distribution of speech signals [103] † simple to adapt [68] † capable to model a sequence of discrete or continuous symbols [13] † inputs can be of variable length [40] † based on the assumption that the probability of being in a particular state is dependent only on its preceding state, ignoring any long-term dependencies [82] † emission probabilities are arbitrarily chosen; hence, these might not even represent properly the output probabilities of the corresponding state [82] ANN (in general) † good classifiers [16, 45] † highly adequate for pattern recognition applications [16, 45] † self-organising [16, 45] † self-learning [16, 45] † self-adaptive in new environments [16, 45] † robust [7] † based on ERM; hence, prone to over training a local minima problems [45, 103] MLP † good discriminating ability [2] † unable to model time distribution of speech signals [2] † inputs have to be of fixed length [2] † able to deal with small vocabularies only [2] SOM † no a priori information is required for training a SOM [134] † can easily adapt if a new sample is presented to it [134] † capable of parallel computation [134] † SOM algorithm is not well defined mathematically; hence, values for the network parameters need to be found by trial-and-error [134] † ordered mapping obtained after the training phase may be lost when applied in real environments due to frequent adaptations [134] RBF † simple to implement [135] † Good discriminating ability [135] † robust [135] † online learning ability [135] † shift invariant in time [91] RNN † able to model time distribution of speech signals thanks to the feedback connections [95, 103] † complex training algorithm [94] † training algorithm is highly sensitive to any changes [94] FNN † does not need large amount of samples during the learning process [99] † ...

show abstract

Section: Automatic Speech Recognition Systemsmentioning

confidence: 99%

Comparative study of automatic speech recognition techniques

Cutajar

Gatt

Grech

et al. 2013

IET signal process.

View full text Add to dashboard Cite

show abstract

“…As a result, the signal is divided into small chunks known as frames, with each frame's signal becoming more stationary and its features represented by a fixed length feature vector ( 11 , 12 ). The procedure is referred to as feature extraction ( 13 , 14 ), and the component that performs it is referred to as the front end. The front end's output is the observation sequence X, with each observation being a feature vector representing a single frame.…”

Section: Introductionmentioning

confidence: 99%

“…The initial stage in speech recognition is the extraction ( 13 , 14 ) of a sequence of feature vectors X that reflects the input voice signal. Feature vectors ( 18 ) are available in a wide range of sizes and shapes.…”

Section: Introductionmentioning

confidence: 99%

Convolutional Neural Network Based Real Time Arabic Speech Recognition to Arabic Braille for Hearing and Visually Impaired

Bhatia

Devi

Alsuwailem

et al. 2022

Front. Public Health

View full text Add to dashboard Cite

Natural Language Processing (NLP) is a group of theoretically inspired computer structures for analyzing and modeling clearly going on texts at one or extra degrees of linguistic evaluation to acquire human-like language processing for quite a few activities and applications. Hearing and visually impaired people are unable to see entirely or have very low vision, as well as being unable to hear completely or having a hard time hearing. It is difficult to get information since both hearing and vision, which are crucial organs for receiving information, are harmed. Hearing and visually impaired people are considered to have a substantial information deficit, as opposed to people who just have one handicap, such as blindness or deafness. Visually and hearing-impaired people who are unable to communicate with the outside world may experience emotional loneliness, which can lead to stress and, in extreme cases, serious mental illness. As a result, overcoming information handicap is a critical issue for visually and hearing-impaired people who want to live active, independent lives in society. The major objective of this study is to recognize Arabic speech in real time and convert it to Arabic text using Convolutional Neural Network-based algorithms before saving it to an SD card. The Arabic text is then translated into Arabic Braille characters, which are then used to control the Braille pattern via a Braille display with a solenoid drive. The Braille lettering triggered on the finger was deciphered by visually and hearing challenged participants who were proficient in Braille reading. The CNN, in combination with the ReLU model learning parameters, is fine-tuned for optimization, resulting in a model training accuracy of 90%. The tuned parameters model's testing results show that adding the ReLU activation function to the CNN model improves recognition accuracy by 84 % when speaking Arabic digits.

show abstract