Sadhana-academy Proceedings in Engineering Sciences, 2021
In this paper, a model is proposed to improve monophone-based connected word speech recognition f... more In this paper, a model is proposed to improve monophone-based connected word speech recognition for the Hindi language by utilizing the Hidden Markov Model (HMM). The model consists of hybrid subword units and domain-specific syntactic structures. The hybrid units contain both phonemeand syllable-based subword units. As the syllable-based subword units cover a larger acoustic span, contextual effects are reduced. The syllable-based acoustic units are applied for modelling only nasal sound in the hybrid model for improving the recognition score of a nasal sound. Further, improvement is proposed using syntactic structures in the grammar definition during the recognition process. Using the domain-specific syntactic structures in the grammar, the search space for the recognizer is reduced; consequently, the performance of the system is improved. For example, two grammar definitions (gram1) with no restriction and grammar(gram2) with domainspecific structures were applied. The speech rec...
Wireless Sensor Network (WSN) is most integral part of our day to day life. They are widely used ... more Wireless Sensor Network (WSN) is most integral part of our day to day life. They are widely used in various applications. Many Clustering based algorithms have been proposed for designing WSN. Some of them are fuzzy based approaches. The purpose of this paper is to propose a clustering based approach with fuzzy logic and relay node implementation that can cope up with drawbacks of the previously proposed algorithms in WSN implementation. Certain parameters like Last Node Die, First Node Die, Half Node Analysis, Cluster Head Energy etc, have been compared with the new proposal. An effort has been made to improve all these parameters so that the network obtained after incorporating it is highly efficient. Improvement has been achieved by using the algorithm in various performance parameters of WSN. The results along with statistics have been clearly shown after implementation of the algorithm.
Digital signal processor, image signal processor and FIR filters have multipliers as an important... more Digital signal processor, image signal processor and FIR filters have multipliers as an important part of their design. On the basis of Vedic mathematics, Vedic multipliers have come out to be very fast multipliers. One of the image processing applications is edge detection. This research presents a small area and high speed 8 bit Vedic multiplier system comprising of compressor based adders. This results in faster edge detection. This architecture is tested on Xilinx vertex 4 FPGA board and simulations were carried out using the Xilinx synthesis tool. Comparisons are made and this system is found to be smaller in area with high speed (the lesser propagation delay). This compressor based Vedic multiplier is 1.1 times speedier than a typical Vedic multiplier. Also, this Vedic Multiplier is 2 times speedier than a ‘simple’ multiplier. Keywords—Detection of edges, Vedic multiplier, image processing, Urdhva Tiryakbhyam sutra.
Wireless Sensor Network is a randomly deployed collection of sensor nodes with the aim to collect... more Wireless Sensor Network is a randomly deployed collection of sensor nodes with the aim to collect the information near its sensing area. The approach is to sense some event, record its respective data value and transmit it to a sink where this data value is utilized and thus become information. The values received by the sink may contain duplicate, inappropriate and inconsistent values. The new research design may focus on collecting and sending only that value which may be utilized by the sink. The transmission of irrelevant data is avoided to increase the performance and lifetime. This paper aims at providing a data management technique which reduces load on sensor nodes to enhance network lifetime. To reduce extra burden on sensor nodes a data prediction model is built which restricts data transmission by predicting future data values. The algorithm finds relationship between data values. The goal is to calculate degree of relatedness between these values so to establish a relati...
This paper presents a robust approach for an automatic speech recognition system (ASR) when both ... more This paper presents a robust approach for an automatic speech recognition system (ASR) when both additive and convolutional noises corrupt the speech signal. Robust features are derived by assuming that the corrupting noise is stationary and the channel effect is fixed during the utterance. In the proposed method the effect of additive and convolutional distortions are minimized by two stage filtering. The first filtering stage includes differential temporal filtering in the autocorrelation domain for reducing additive noise effects, followed by additional filtering in the logarithmic spectrum domain to reduce convolutional noise effects. Convolutional channel distortion is assumed to be linear and time invariant. A task of multispeaker isolated Hindi word recognition is conducted to demonstrate the effectiveness of using these robust features. The cases of channel filtered speech signal corrupted by white noise and different colored noises such as factory, babble and F16, which are...
This research paper presents highly optimized barrel shifter at 22nm Hi K metal gate strained Si ... more This research paper presents highly optimized barrel shifter at 22nm Hi K metal gate strained Si technology node. This barrel shifter is having a unique combination of static and dynamic body bias which gives lowest power delay product. This power delay product is compared with the same circuit at same technology node with static forward biasing at ‘supply/2’ and also with normal reverse substrate biasing and still found to be the lowest. The power delay product of this barrel sifter is .39362X10-17J and is lowered by approximately 78% to reference proposed barrel shifter at 32nm bulk CMOS technology. Power delay product of barrel shifter at 22nm Hi K Metal gate technology with normal reverse substrate bias is 2.97186933X10-17J and can be compared with this design’s PDP of .39362X10-17J. This design uses both static and dynamic substrate biasing and also has approximately 96% lower power delay product compared to only forward body biased at half of supply voltage. The NMOS model use...
After the reliable commercial use of speech based processing be it text to speech conversion vice... more After the reliable commercial use of speech based processing be it text to speech conversion vice-versa, Interactive response system, product marketing or speech based search. Researchers are now focused on to the study of the emotional content of speech signals, and hence, many systems have been proposed for the classification and identification of the emotional content of a spoken utterance. In this paper attempt are made to present the features of emotional Hindi language, for identification and classification. Speech databases have been used for speech recognition, speech synthesis, speaker recognition, language translation, emotion recognition, emotion conversion. An attempt has been made classify emotional speech corpus developed at IPEC. A study is also preformed to understand the effect of feature selection in emotion identification and classification. Emotional speech classification is preformed on the basis of LPCC and MFCC features are also performed using K-mean cluster ...
The research paper describes the implementation of different machine learning algorithms used for... more The research paper describes the implementation of different machine learning algorithms used for classification. The aim is to provide the interested learner about the implementation of the classification algorithms on the real-world earthquake dataset. This study has been done using support vector machine (SVM), K-nearest neighbor (KNN), random forest (RF), and Naive Bayes (NB) algorithms in R programming language. The result of the study is the analysis and classification of each data value in the dataset and hence assigning it to the correct class label. The study is done on the dataset collected from the Web site of Indian Meteorological Department, Ministry of Earth Sciences, Government of India. The methodology includes data acquisition and description, feature selection, data normalization, data partitioning, model implementation and prediction, optimization and fine-tuning, classification and finding rate of accuracy, and misclassification error. The research work also comp...
Opportunistic Networks can be defined as Delay Tolerant Network, which are formed dynamically wit... more Opportunistic Networks can be defined as Delay Tolerant Network, which are formed dynamically with participating nodes’ help. Opportunistic Networks follows Store-Carry-Forward principle to deliver/route the data in the network. Routing in Opportunistic Network starts with the Seed Node (Source Node) which delivers the data with the help of Intermediate nodes. Intermediate nodes store the data while roaming in the network until it comes in contact with appropriate forwarding node (relay node) or destination node itself. An extensive literature survey is performed to analyse various routing protocols defined for Opportunistic Network. With mobility induced routing, establishing and maintaining the routing path is a major challenge. Further, Store-Carry-Forward routing paradigm imposes various challenges while implementing and executing the network. Due to the unavailability of the suitable relay node, data needs to be stored within the Node’s Memory, imposes buffer storage issues at ...
International Journal of Ambient Computing and Intelligence
Recent developments in information gathering procedures and the collection of big data over a per... more Recent developments in information gathering procedures and the collection of big data over a period of time as a result of introducing high computing devices pose new challenges in sensor networks. Data prediction has emerged as a key area of research to reduce transmission cost acting as principle analytic tool. The transformation of huge amount of data into an equivalent reduced dataset and maintaining data accuracy and integrity is the prerequisite of any sensor network application. To overcome these challenges, a data prediction technique is suggested to reduce transmission of redundant data by developing a regression model on linear descriptors on continuous sensed data values. The proposed model addresses the basic issues involved in data aggregation. It uses a buffer based linear filter algorithm which compares all incoming values and establishes a correlation between them. The cluster head is accountable for predicting data values in the same time slot, calculates the devia...
This correspondence deals with the segmentation of a video clip into independently moving visual ... more This correspondence deals with the segmentation of a video clip into independently moving visual objects. This is an important step in structuring video data for storage in digital libraries. The method follows a bottom-up approach. The major contribution is a new well-...
Sadhana-academy Proceedings in Engineering Sciences, 2021
In this paper, a model is proposed to improve monophone-based connected word speech recognition f... more In this paper, a model is proposed to improve monophone-based connected word speech recognition for the Hindi language by utilizing the Hidden Markov Model (HMM). The model consists of hybrid subword units and domain-specific syntactic structures. The hybrid units contain both phonemeand syllable-based subword units. As the syllable-based subword units cover a larger acoustic span, contextual effects are reduced. The syllable-based acoustic units are applied for modelling only nasal sound in the hybrid model for improving the recognition score of a nasal sound. Further, improvement is proposed using syntactic structures in the grammar definition during the recognition process. Using the domain-specific syntactic structures in the grammar, the search space for the recognizer is reduced; consequently, the performance of the system is improved. For example, two grammar definitions (gram1) with no restriction and grammar(gram2) with domainspecific structures were applied. The speech rec...
Wireless Sensor Network (WSN) is most integral part of our day to day life. They are widely used ... more Wireless Sensor Network (WSN) is most integral part of our day to day life. They are widely used in various applications. Many Clustering based algorithms have been proposed for designing WSN. Some of them are fuzzy based approaches. The purpose of this paper is to propose a clustering based approach with fuzzy logic and relay node implementation that can cope up with drawbacks of the previously proposed algorithms in WSN implementation. Certain parameters like Last Node Die, First Node Die, Half Node Analysis, Cluster Head Energy etc, have been compared with the new proposal. An effort has been made to improve all these parameters so that the network obtained after incorporating it is highly efficient. Improvement has been achieved by using the algorithm in various performance parameters of WSN. The results along with statistics have been clearly shown after implementation of the algorithm.
Digital signal processor, image signal processor and FIR filters have multipliers as an important... more Digital signal processor, image signal processor and FIR filters have multipliers as an important part of their design. On the basis of Vedic mathematics, Vedic multipliers have come out to be very fast multipliers. One of the image processing applications is edge detection. This research presents a small area and high speed 8 bit Vedic multiplier system comprising of compressor based adders. This results in faster edge detection. This architecture is tested on Xilinx vertex 4 FPGA board and simulations were carried out using the Xilinx synthesis tool. Comparisons are made and this system is found to be smaller in area with high speed (the lesser propagation delay). This compressor based Vedic multiplier is 1.1 times speedier than a typical Vedic multiplier. Also, this Vedic Multiplier is 2 times speedier than a ‘simple’ multiplier. Keywords—Detection of edges, Vedic multiplier, image processing, Urdhva Tiryakbhyam sutra.
Wireless Sensor Network is a randomly deployed collection of sensor nodes with the aim to collect... more Wireless Sensor Network is a randomly deployed collection of sensor nodes with the aim to collect the information near its sensing area. The approach is to sense some event, record its respective data value and transmit it to a sink where this data value is utilized and thus become information. The values received by the sink may contain duplicate, inappropriate and inconsistent values. The new research design may focus on collecting and sending only that value which may be utilized by the sink. The transmission of irrelevant data is avoided to increase the performance and lifetime. This paper aims at providing a data management technique which reduces load on sensor nodes to enhance network lifetime. To reduce extra burden on sensor nodes a data prediction model is built which restricts data transmission by predicting future data values. The algorithm finds relationship between data values. The goal is to calculate degree of relatedness between these values so to establish a relati...
This paper presents a robust approach for an automatic speech recognition system (ASR) when both ... more This paper presents a robust approach for an automatic speech recognition system (ASR) when both additive and convolutional noises corrupt the speech signal. Robust features are derived by assuming that the corrupting noise is stationary and the channel effect is fixed during the utterance. In the proposed method the effect of additive and convolutional distortions are minimized by two stage filtering. The first filtering stage includes differential temporal filtering in the autocorrelation domain for reducing additive noise effects, followed by additional filtering in the logarithmic spectrum domain to reduce convolutional noise effects. Convolutional channel distortion is assumed to be linear and time invariant. A task of multispeaker isolated Hindi word recognition is conducted to demonstrate the effectiveness of using these robust features. The cases of channel filtered speech signal corrupted by white noise and different colored noises such as factory, babble and F16, which are...
This research paper presents highly optimized barrel shifter at 22nm Hi K metal gate strained Si ... more This research paper presents highly optimized barrel shifter at 22nm Hi K metal gate strained Si technology node. This barrel shifter is having a unique combination of static and dynamic body bias which gives lowest power delay product. This power delay product is compared with the same circuit at same technology node with static forward biasing at ‘supply/2’ and also with normal reverse substrate biasing and still found to be the lowest. The power delay product of this barrel sifter is .39362X10-17J and is lowered by approximately 78% to reference proposed barrel shifter at 32nm bulk CMOS technology. Power delay product of barrel shifter at 22nm Hi K Metal gate technology with normal reverse substrate bias is 2.97186933X10-17J and can be compared with this design’s PDP of .39362X10-17J. This design uses both static and dynamic substrate biasing and also has approximately 96% lower power delay product compared to only forward body biased at half of supply voltage. The NMOS model use...
After the reliable commercial use of speech based processing be it text to speech conversion vice... more After the reliable commercial use of speech based processing be it text to speech conversion vice-versa, Interactive response system, product marketing or speech based search. Researchers are now focused on to the study of the emotional content of speech signals, and hence, many systems have been proposed for the classification and identification of the emotional content of a spoken utterance. In this paper attempt are made to present the features of emotional Hindi language, for identification and classification. Speech databases have been used for speech recognition, speech synthesis, speaker recognition, language translation, emotion recognition, emotion conversion. An attempt has been made classify emotional speech corpus developed at IPEC. A study is also preformed to understand the effect of feature selection in emotion identification and classification. Emotional speech classification is preformed on the basis of LPCC and MFCC features are also performed using K-mean cluster ...
The research paper describes the implementation of different machine learning algorithms used for... more The research paper describes the implementation of different machine learning algorithms used for classification. The aim is to provide the interested learner about the implementation of the classification algorithms on the real-world earthquake dataset. This study has been done using support vector machine (SVM), K-nearest neighbor (KNN), random forest (RF), and Naive Bayes (NB) algorithms in R programming language. The result of the study is the analysis and classification of each data value in the dataset and hence assigning it to the correct class label. The study is done on the dataset collected from the Web site of Indian Meteorological Department, Ministry of Earth Sciences, Government of India. The methodology includes data acquisition and description, feature selection, data normalization, data partitioning, model implementation and prediction, optimization and fine-tuning, classification and finding rate of accuracy, and misclassification error. The research work also comp...
Opportunistic Networks can be defined as Delay Tolerant Network, which are formed dynamically wit... more Opportunistic Networks can be defined as Delay Tolerant Network, which are formed dynamically with participating nodes’ help. Opportunistic Networks follows Store-Carry-Forward principle to deliver/route the data in the network. Routing in Opportunistic Network starts with the Seed Node (Source Node) which delivers the data with the help of Intermediate nodes. Intermediate nodes store the data while roaming in the network until it comes in contact with appropriate forwarding node (relay node) or destination node itself. An extensive literature survey is performed to analyse various routing protocols defined for Opportunistic Network. With mobility induced routing, establishing and maintaining the routing path is a major challenge. Further, Store-Carry-Forward routing paradigm imposes various challenges while implementing and executing the network. Due to the unavailability of the suitable relay node, data needs to be stored within the Node’s Memory, imposes buffer storage issues at ...
International Journal of Ambient Computing and Intelligence
Recent developments in information gathering procedures and the collection of big data over a per... more Recent developments in information gathering procedures and the collection of big data over a period of time as a result of introducing high computing devices pose new challenges in sensor networks. Data prediction has emerged as a key area of research to reduce transmission cost acting as principle analytic tool. The transformation of huge amount of data into an equivalent reduced dataset and maintaining data accuracy and integrity is the prerequisite of any sensor network application. To overcome these challenges, a data prediction technique is suggested to reduce transmission of redundant data by developing a regression model on linear descriptors on continuous sensed data values. The proposed model addresses the basic issues involved in data aggregation. It uses a buffer based linear filter algorithm which compares all incoming values and establishes a correlation between them. The cluster head is accountable for predicting data values in the same time slot, calculates the devia...
This correspondence deals with the segmentation of a video clip into independently moving visual ... more This correspondence deals with the segmentation of a video clip into independently moving visual objects. This is an important step in structuring video data for storage in digital libraries. The method follows a bottom-up approach. The major contribution is a new well-...
Uploads
Papers by Amita Dev