%0 Journal Article %1 Gaikwad_2015 %A Gaikwad, Santosh %A Gawali, Bharti %A Mehrotra, Suresh %D 2015 %I Auricle Technologies, Pvt., Ltd. %J International Journal on Recent and Innovation Trends in Computing and Communication %K Appliances Electrical FAR FRR HM200 Human Interaction Machine Recognition Responding Speech Time WER %N 4 %P 2273--2281 %R 10.17762/ijritcc2321-8169.1504110 %T Marathi Speech Interface System for the Activation and Controlling of Electronic Equipment %U http://dx.doi.org/10.17762/ijritcc2321-8169.1504110 %V 3 %X This paper provides a framework to design Marathi speech recognition interface system to control the electronic equipment like fan and electric bulb. This system is designed using HM2007 speech recognition kit. HM2007 is a single chip CMOS LSI circuit with on chip front end and analysis. The interface system is trained with the database of total 15 most commonly used words to activate and control the equipment. The training and testing of Marathi speech interface system is done in noisy and noise free environment. The performance of MSAEE system calculated on the basis of False Acceptance Rate (FAR), False Recognition Rate (FRR), Word Error Rate (WER), sensitivity, specificity and F-Measure. The average accuracy obtained in noisy environment is 88.77% whereas the accuracy of 98.88% is achieved without noise. The sensitivity value is superior then specificity means MSAEE system is robust and dynamic. FAR values is most significant than value of FRR so, it means MSAEE system proved a best solution for controlling electrical appliances. HM2007 is found to be most promising speech recognition IC, through which many speech interface application can be built that will be beneficial to people with less mobility.

@article{Gaikwad_2015, abstract = {This paper provides a framework to design Marathi speech recognition interface system to control the electronic equipment like fan and electric bulb. This system is designed using HM2007 speech recognition kit. HM2007 is a single chip CMOS LSI circuit with on chip front end and analysis. The interface system is trained with the database of total 15 most commonly used words to activate and control the equipment. The training and testing of Marathi speech interface system is done in noisy and noise free environment. The performance of MSAEE system calculated on the basis of False Acceptance Rate (FAR), False Recognition Rate (FRR), Word Error Rate (WER), sensitivity, specificity and F-Measure. The average accuracy obtained in noisy environment is 88.77% whereas the accuracy of 98.88% is achieved without noise. The sensitivity value is superior then specificity means MSAEE system is robust and dynamic. FAR values is most significant than value of FRR so, it means MSAEE system proved a best solution for controlling electrical appliances. HM2007 is found to be most promising speech recognition IC, through which many speech interface application can be built that will be beneficial to people with less mobility.}, added-at = {2015-08-26T10:23:53.000+0200}, author = {Gaikwad, Santosh and Gawali, Bharti and Mehrotra, Suresh}, biburl = {https://www.bibsonomy.org/bibtex/23b23999fc187bdfe8ec8f35a5a660fb7/ijritcc}, doi = {10.17762/ijritcc2321-8169.1504110}, interhash = {df9bbfad82cf5ccfef001274a4507a49}, intrahash = {3b23999fc187bdfe8ec8f35a5a660fb7}, journal = {International Journal on Recent and Innovation Trends in Computing and Communication}, keywords = {Appliances Electrical FAR FRR HM200 Human Interaction Machine Recognition Responding Speech Time WER}, month = {april}, number = 4, pages = {2273--2281}, publisher = {Auricle Technologies, Pvt., Ltd.}, timestamp = {2015-08-26T10:23:53.000+0200}, title = {Marathi Speech Interface System for the Activation and Controlling of Electronic Equipment}, url = {http://dx.doi.org/10.17762/ijritcc2321-8169.1504110}, volume = 3, year = 2015 }