Computer Vision

The first project I made in college that got me into Computer Vision and Machine Learning. By tracking the end of a retroreflector-tipped harry potter wand replica and training my own SVM for handwritten character recognition, I was able to do magic in real life. This was demostrated by the spell Alohomora that opens a box when you write ‘A’ in mid-air using the wand. The project got featured in official blogs at Raspberry Pi, Hackaday and a complete article in the 86th issue of MagPi, the official Raspberry Pi magazine

First prize winning project for Mitsubishi America’s Rally-To-Tally Online Hackathon. Developed an autonomous solution for their problem statement of inventory management of large industrial steel pipes using computer vision and outdoor mobile robots. Demostrated the PoC with the help of a custom simulation of Husky robot in Gazebo with InfluxDB and Grafana for integration for the data visualization