Posts by Collection





Lectured on various topics in Deep Learning

Lab Talk, IIT Madras, EE Dept., 2015

This Lecture Series started at a time when Deep Learning was taking off, and it aims to explain the (then) emerging field of Deep Learning for the (then) members of Image Processing and Computer Vision Lab and Computational Imaging Lab, headed by Prof. A. N. Rajagopalan and Prof. Kaushik Mitra, respectively.
                Here, the first lecture introduces Deep Learning (L1), later, followed by a doubt-clearing session (L1.1). Lecture 2 covers Convolutional Neural Network (L2), and Lecture 3 provides a mathematical interpretation of Deep Learning; O Lost! This series is concluded with Lecture 4, which discusses several Deep Learning approaches for Image Processing. [Slides-L1] [Slides-L1.1] [Slides-L2] [Slides-L3] [Slides-L4]

Mentored two undergrad students towards the best Deep Learning course-project

Graduate Course (Deep Learning), IITM, EE Dept., 2017

Here, I proposed a problem statement Efficient Sketch Classification for Resource-constrained platforms, and designed a research plan for the same. Undergrads Akshit Kumar and Sachin Agrawal joined with me and finally got recognition for the best Deep Learning course-project. Presently, both the students are continuing research in Deep Learning; the former is a Master's student at the University of Michigan, US, and the latter is a Data Scientist at Daimler, Germany. [Slides]

Mentored (with Prof. A. N. Rajagopalan) a project associate towards a funded project.

Funded Project, IITM, EE Dept., 2018

Lots of accidents happen while exiting a vehicle from its side doors, for it is difficult to ensure whether another vehicle(s) are coming from behind and how fast it is. This project attempts to solve this issue; project associate Madumithra Krishnamoorthy was incharge. Together, we employ a side view camera and propose a deployable algorithm using image processing techniques to address this. Later, she joined as a graduate research student in image processing. [Demo]