Mini Projects
Some of these project have been self supervised.
Kaggle Competition: CDiscount challenge Image Classification
Course Project (October, 2017 - November, 2017) at York University
Collaborators:
Mr. Robert Codd Downey, Department of Electrical Engineering and Computer Science, York University
Prof. Manos Papagelis, Department of Electrical Engineering and Computer Science, York University
We participated in a Kaggle competition based on image classification. There are 5270 different categories and over 12 million images to train from. Intelligent Data pre-processing techniques and Deep Learning based models were used using tensor flow.
Collaborators:
Mr. Robert Codd Downey, Department of Electrical Engineering and Computer Science, York University
Prof. Manos Papagelis, Department of Electrical Engineering and Computer Science, York University
We participated in a Kaggle competition based on image classification. There are 5270 different categories and over 12 million images to train from. Intelligent Data pre-processing techniques and Deep Learning based models were used using tensor flow.
Reverse Engineering the Oculus Prime SLAM Navigator
Self Supervised (November, 2016 - December 2016)
Part-time hobby
Currently I am using the Oculus Prime SLAM Navigator robot to fully understand the architecture of navigation and SLAM algorithms being used by xaxxon. After the extensively using the robot for building an accurate map based on vision in my apartment environment, I will be using the robot on a daily bases for my personal uses like: waking me up in the morning (acting as a motion based alarm), surveillance at my house remotely when I am working elsewhere, object following behaviour, etc.
Part-time hobby
Currently I am using the Oculus Prime SLAM Navigator robot to fully understand the architecture of navigation and SLAM algorithms being used by xaxxon. After the extensively using the robot for building an accurate map based on vision in my apartment environment, I will be using the robot on a daily bases for my personal uses like: waking me up in the morning (acting as a motion based alarm), surveillance at my house remotely when I am working elsewhere, object following behaviour, etc.
Content Based Image retrieval based Search Engine for personalized search on computers
Online Supervision through Skype (September, 2014)
Self supervised later (December 2014)
Collaborators:
Online Collaboration - Dr. Visvanathan Ramesh, Professor J.W. Goethe University, Frankfurt, Germany
Searching for specific images in our laptops / computers can be a difficult task. Here we propose to develop a system where we incorporate the concept of content based image retrieval. Suppose you want to search for an image of yourself in a forested area. So you enter in our software that you want all images in your computer which indicate the presence of both 'a forested area' and 'your face' in that area. Alternatively if you want to find an image where you are present with your friends near suppose a lake. So our software would take as 3 input parameters - 'your face', 'humans' and 'water body'. It then shows the desired result.
* will be working on this project later this year
Self supervised later (December 2014)
Collaborators:
Online Collaboration - Dr. Visvanathan Ramesh, Professor J.W. Goethe University, Frankfurt, Germany
Searching for specific images in our laptops / computers can be a difficult task. Here we propose to develop a system where we incorporate the concept of content based image retrieval. Suppose you want to search for an image of yourself in a forested area. So you enter in our software that you want all images in your computer which indicate the presence of both 'a forested area' and 'your face' in that area. Alternatively if you want to find an image where you are present with your friends near suppose a lake. So our software would take as 3 input parameters - 'your face', 'humans' and 'water body'. It then shows the desired result.
* will be working on this project later this year
Student Mentorship Program
Mentor for teaching Machine Learning (August, 2013 - April, 2014)
Through this program, I taught a group of 12 sophomore students the basic concepts of Machine Learning. I took classes once a week for 8 months. At the end of the program, I supervised 5 basic projects which my mentees finished.
Compiler Created for the language C# (March 2014 - April 2014)
under guidance of Professor Aruna Malapati
A basic compiler for the language C# was created. Lexer and Parser was created after building the appropriate grammar for the language. Coding was entirely done in C.
Click here for code.
DNS Client Application and DNS Client-Server (October, 2013 - November 2013)
under the guidance of Mr. Gokaul Kanan Sadasivam, Lecturer BITS Pilani Hyderabad Campus
A DNS client application where in client can perform a number of operations by sending messages to the internet was develpoed. Based on a given domain name, it performed a number of functions such as retrieving all IP addresses, mail server records, host info, etc. The later part of the project dealt with implementing iterative and recursive queries by.
click here to download report
Robot which selectively removes impurities in the form of objects (August 2013)
Kicker mechanism which removes blocks based on their colors and shapes to selectively filter out the objects running on a conveyor belt using image processing techniques.
click here for code
Games in Python (July 2013)
Created games in python like blackjack, asteroids, pong, memory game, pacman, spaceships etc. as a part of the online course on Interactive programming in Python by Rice University
click here to download games code
Games in Java (December 2012 - January 2013)
Created games in Java like memory game, snakes, pacman, etc. as a part of the winter internship in a start-up company experts-hub
GPS Tracker to map a given area in real time (January 2013)
MATLAB based application to generate on map the path traveled. A GPS USB module was used. A distance of around 50 kilometers was traveled in car and at the end the code written in MATLAB mapped the actual path traveled
Maze Solver Robot (October 2012)
Maze Solving bot was created to solve any given maze. PID algorithm was used for line following and a right first algorithm was used for assigning priorities to various directions for scanning the maze
PID Algorithm based 6 sensor Line Follower (October, 2011)
A 6 sensor PID algorithm implementing line follower was designed as a project kit for Valonia and over 300 kits were sold through the course of our workshops
Mentor for teaching Machine Learning (August, 2013 - April, 2014)
Through this program, I taught a group of 12 sophomore students the basic concepts of Machine Learning. I took classes once a week for 8 months. At the end of the program, I supervised 5 basic projects which my mentees finished.
- Hand Written Digit Recognition
- Gesture Controlled Game
- Sign Language Recognition
- Image Compression
- Spam Filter
Compiler Created for the language C# (March 2014 - April 2014)
under guidance of Professor Aruna Malapati
A basic compiler for the language C# was created. Lexer and Parser was created after building the appropriate grammar for the language. Coding was entirely done in C.
Click here for code.
DNS Client Application and DNS Client-Server (October, 2013 - November 2013)
under the guidance of Mr. Gokaul Kanan Sadasivam, Lecturer BITS Pilani Hyderabad Campus
A DNS client application where in client can perform a number of operations by sending messages to the internet was develpoed. Based on a given domain name, it performed a number of functions such as retrieving all IP addresses, mail server records, host info, etc. The later part of the project dealt with implementing iterative and recursive queries by.
click here to download report
Robot which selectively removes impurities in the form of objects (August 2013)
Kicker mechanism which removes blocks based on their colors and shapes to selectively filter out the objects running on a conveyor belt using image processing techniques.
click here for code
Games in Python (July 2013)
Created games in python like blackjack, asteroids, pong, memory game, pacman, spaceships etc. as a part of the online course on Interactive programming in Python by Rice University
click here to download games code
Games in Java (December 2012 - January 2013)
Created games in Java like memory game, snakes, pacman, etc. as a part of the winter internship in a start-up company experts-hub
GPS Tracker to map a given area in real time (January 2013)
MATLAB based application to generate on map the path traveled. A GPS USB module was used. A distance of around 50 kilometers was traveled in car and at the end the code written in MATLAB mapped the actual path traveled
Maze Solver Robot (October 2012)
Maze Solving bot was created to solve any given maze. PID algorithm was used for line following and a right first algorithm was used for assigning priorities to various directions for scanning the maze
PID Algorithm based 6 sensor Line Follower (October, 2011)
A 6 sensor PID algorithm implementing line follower was designed as a project kit for Valonia and over 300 kits were sold through the course of our workshops