http://103.99.128.19:8080/xmlui/handle/123456789/127 Proceedings published in the Dept. of Physics 2026-04-19T09:39:01Z http://103.99.128.19:8080/xmlui/handle/123456789/130 Title: Voice Command Based Matlab GUI for Microcontroller Authors: Dept. of Physics, Dept. of Physics Abstract: Abstract—Voice command based applications use speech to make people’s life more easy and flexible. It is one of the most interesting topics in recent research spheres and still there are lots of scope for improvement. The paper represents the method of controlling electrical components by Matlab graphical user interface (GUI) using voice command. The idea of developing voice controlled application can be established through speech identification and speech verification. Speech identification is performed by feature extracting of speech signal utilizing Mel Frequency Cepstrum Coefficient (MFCC) technique in Matlab. Then vector quantization (VQ) method has been proposed for verification. The recognized speech will be compared with prerecorded database and according to those results the desired commandsaresentfromapplicationtomicrocontrollertocontrol the devices. 2018-06-28T00:00:00Z http://103.99.128.19:8080/xmlui/handle/123456789/129 Title: ECG signal Compression using Data Extraction and Truncated Singular Value Decomposition Authors: Dept. of Physics, Dept. of Physics Abstract: Abstract—In this paper a novel algorithm is used for compressing ECG data. The algorithm uses two techniques simultaneously to get a better compression of ECG data. Firstly Data extraction: a sampling technique based on a user defined threshold value is applied on ECG followed by Truncated Singular Value Decomposition. Changing the number of Singular Values higher compression can be achieved. The proposed method provides two variables: Threshold & Number of Singular Value, which can be changed to achieve optimum compression while keeping the error low. Compression performance & quality of reconstructed ECG in the proposed method is evaluated using Compression Ratio (CR) & Percent Root Mean Square Difference (PRD) 2018-06-28T00:00:00Z http://103.99.128.19:8080/xmlui/handle/123456789/128 Title: A Communication Platform Between Bangla and Sign Language Authors: Dept. of Physics, Dept. of Physics Abstract: Abstract—Hearing and speaking impairments are very crippling handicaps for human beings. People affected by such impairments often use sign language to communicate, which is very different from spoken and written language, often leading to difficulties in communication when a translation medium is not readily available. This leads to heavy struggle in the daily lives of the affected people, especially in Bangladesh. A simple, accurate and portable translation medium is necessary to address this problem, which would allow two-way communication between a normal person and an impaired person. Speech recognition and conversion to sign language is a key part of such a medium, especially since Bangla is a rather challenging language to work with. In this paper, we have described the development of a smartphone application, which provides a two-way communication platform, which employs a speech-to-sign language system and a text-to-speech system. Testing of the application shows satisfactory results, with an average accuracy of 84.71% in various test environments. 2018-06-28T00:00:00Z