Farhan Ar Rafi

Research

I started my research during my undergraduate thesis. During undergraduate we extended Geonames to create Bengali Geonames. The target was to create Bengali Geowordnet so that we could relate one place to another. I have listed all my research below including the one I talked about. They are in no specific order.
image

Application of Android and Raspberry Pi for Home Automation

March 201810 min read
Home Automation can be used to automate various activities in home and office by interfacing with traditional electrical and mechanical devices. Using Raspberry Pi for home automation has numerous advantages such as low power consumption, low cost and powerful hardware with a very small footprint. In this project we are proposing a system to intuitively control and monitor electrical components at home or office using a real time control system using Raspberry Pi and Android. The proposed system can effectively reduce waste of energy, save cost and add convenience to daily life. The system uses an Android application as the user interface and the Raspberry PI is used to control and monitor electrical components via GPIO pins. The system is inexpensive, secure, easy to implement and uses open-source technologies.
Read more →
image

Geonames Classification for Bengali Geowordnet

December 201515 min read
Bengali Geo-Wordnet focuses on the creation of system that integrates Wordnet, Geonames and Bengali Language that provides a growing vocabulary set and crowd sourced data for all linguistic and natural language processing systems. The Geonames data used in the system has been prepared using manual classifications methods under the supervision of top linguists and professors of related departments. Validation data shows that 98.8% terms were classified properly with 1.2% being non-existent in Bengali language. The underlying system has been designed using a scalable method which enables the system to be extended and replicated with ease. The system opens wide possibilities of development of Bengali Linguistic Systems and Natural Language Processing Systems and Optical Character Recognition Systems.
Read more →
image

Real-Time Human Pose Detection and Body Shape Estimation For Virtual Fitting of Cloths

Before 20153 min read
The main focus of the research is to generate a accurate,dynamic 3D image of a human subject in real-time and without processing huge data sets and having stored data of the subject. The purposeof this research is to find a solution to the problem of acquiring bodyshape and detecting posture of human subjects. Using multiple Kinect cameras, human body pose and shape is scanned and the data is used to reconstruct the human body shape. Shape and pose can be obtained from both moving and static objects. So these data can be used to produce an interactive real-time solution for virtual fitting of clothings.
Read more →
Choose Colour