Hello, I'm Shuvo Kumar Paul from Dhaka, Bangladesh. I've completed my BSc in CSE from North South University and earned my MS in CS degree from Independent University, Bangladesh. I've worked both as a researcher and as a Soft./Web developer. I'm currently working as a Research Associate at the Computer Vision & Cybernetics group under the supervision of Dr. M. Ashraful Amin. My research interests include, but are not limited to, Computer Vision, Machine Learning, Deep Learning, Brain Computer Interface, Human Computer Interaction, and Natural Language Processing.
Research Area: Natural Language Processing, Deep Learning, Human Computer Interaction
Currently working on "Bengali speech recognition system" and "key word based image description and information retrieval".
While working as a System Engineer at FAT I worked daily to employ my programming skills in the development of a number of technical projects, web solutions and assignments which help the monitoring process of various telecom tasks of Grameen Phone.
Developed one-click solutions for various telecom configuration tasks.
Built a web task management system.
Developed automated solutions for various integration tasks.
Research Area: Computer Vision, Machine Learning, Brain Computer Interface
Developed a Data Acquisition tool in python for the MUSE brain sensing headband.
Worked on gesture recognition with the help of Microsoft Kinect.
Developed web solutions using Joomla, Laravel, Wordpress.
Abstract: Human brain uses a complex electro-chemical signaling pattern that creates our imagination, memory and self-consciousness. It is said that Electroencephalography better known as EEG contains signatures of various tasks that we perform. In this paper we study the possibility of categorizing tasks conducted by humans from EEG recordings. The novelty of this study mainly lies in the use of very cost effective consumer grade wireless EEG devices. Three cognitive tasks were considered: text reading and writing, Math problem solving and watching videos. Twelve subjects were used in this experiment. Initial features were calculated from Discrete Wavelet Transform (DWT) of raw EEG signals. After application of appropriate dimensionality reduction, Support Vector Machine (SVM) was used for classification of tasks. DWT + Kernel PCA with SVM based classifier showed 86.09 % accuracy.
Abstract: This work reports the design, construction and control of a two-wheel object tracking robot. The Android controlled Arduino bot is a robot armed with a brain to recognize and track objects persuasively. Object recognition and tracking were achieved through extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object. However, we have also demonstrated tracking through offline data because of the limitations of the processing power of the cellular devices.
Developed a windows app called "Project Beetle" that had a suite of tools capable of diagnosing, tracking, and predicting crop disease.
Was awarded a scholarship based on admission test and previous academic achievements.
Was awarded a scholarship based on previous research work and academic achievements.
Have been organizing charity fund raising and winter clothing drive every year
Participated in various inter-cultural activities organized by Confucius Institute, NSU.
Have been an active theater activist since 2015.