CS698U - Topics in Computer Vision

Instructor: Gaurav Sharma, Assistant Professor, CSE, IITK
Contact: grv AT cse.iitk.ac.in ([CS698U] in email subject; o/w ignored)
Lectures: Mon and Wed, 17h00–18h30
Office hours: By appointment on email
TAs: Aishwarya Gupta (aish), Ayushman Singh Sisodiya (ayushmn), Gundeep Arora (gundeep), Pawan Kuman (kpawan)
TA emails: username@iitk.ac.in

Summary

The course will make the students familiar with basics of learning-based as well as geometric computer vision. The list of possible topics will be

• Convolutional Neural Networks
• Recurrent Neural Networks
• Autoencoders
• Camera calibration
• Epipolar geometry
• 3D reconstruction

This list will evolve based on the level of students enrolled and their interests. For each of the topics we will start with the basics, touch upon some current applications and then you would be expected to work on an assignment which would have a strong programming component. The course is expected to give you a good foundation if you would like to work on Computer Vision in the future either in academic or industrial research and development.

Department's anti-cheating policy is applicable on your participation in the course.

Grading

• Assignments 40%
• Quizes 20%
• Mid-sem 15%
• End-sem 25%

Announcements

• Assignment 1 released; deadline is 31 Jan, Tue, 23h59
• Assignment 2 (task) released; deadline is 12 Feb, Sun, 23h59
• Assignment 3 released; deadline is 9 Mar 22 Mar, Wed, 16h00
• Assignment 4 released; deadline is 16 Apr 19 Apr, Wed, 23h59

Assignments

Please use the IITK CSE moodle website for submitting the assignments. The time stamp of the moodle upload will be considered your submission time.
Any submission by mail will not be considered.

1. Assignment 1: Implement a Multi Layer Perceptron classifier for handwritten digit classification. The PDF of the problem statement has been circulated and put on Moodle as well. Submission deadline is 31 Jan, Tue, 23h59
2. Assignment 3: Implement a LeNet-5; details sent via mailing list and put on Moodle. Submission deadline is 22 Mar, Wed, 16h00
3. Assignment 4: Remove perspective distortion using DLT and Panorama stitching using robust Homography estimation. Submission deadline is 19 Apr, Wed, 11h59

Late Submission Policy

You have a total of 4 late days without penalty, i.e. you could be late by 4 days for A1 or 3 for A1 and 1 for A2 and so on.

Course Calendar

All deadlines below are at 23:59 IST of the respective dates.
The course material (slides etc.) are freely usable for educational and non-commercial research purpose, with due attribution. The material is avialable as is without any warranty, expressed or implied, whatsoever. Any commercial use requires prior written permission from the author. If you are the owner of any of the content included (eg. images), and feel that it has been unfairly used, kindly let me know and I will either attribute it to you as you specify or take it off, depending on your request.

Related Courses, Blogs

• CS231n, CNN for Visual Recognition, Course, Stanford
• An Intuitive Explanation of Convolutional Neural Networks
• Recurrent Neural Networks Tutorial
• The Unreasonable Effectiveness of Recurrent Neural Networks
• Understanding LSTM Networks
• Mark Neilson’s Book
• Iamtrask’s coding blog