This repository contains the assignments of Computer Vision course taught by Dr. Soriyani at IUST
You can find explanation in each folder. In this section, there is just summary of each work folder:
This work consists of two sections:
- Difference between gaussian and average filter for smoothing image
- Using multiple gaussian filters on single image consecutively and the relation between the results
This work consists of using Fourier transform to extract the phase and amplitude of two different images then:
- Combining the phase of image 1 and amplitude of image 2 and vice versa
- Using high-pass and low-pass filters on image 1 and image 2 respectively
- Combining these results w.r.t. different cut-off values
- Demonstrating that phase matters not amplitude!
In this work, texture classification has been discussed. You can find following contents:
- CUReT dataset
- Classification w/o filters
- Classification with filters
- MR8 filter banks
- Results
This work contains information about using different topics such as:
- Canny Edge Detector
- Gaussian Noise
- Gradient Intensification
- Non-Max Suppression
- Thresholding
- Straight Line Detector
- Assign each edge to a direction
- Getting edgelets using connected components
- Compute straightness and direction
- Threshold
In this work, following topics has been covered:
- Computing homography matrix of two images
- Using computed homography to warp source image to the destination image
- Using two previous steps on a video to create VR
I have used many contents from different websites which I have not referenced! I have bookmarked all those references and I will update the slides or at least provide the links at each folder.
The instructor has been used slides and videos of Derek Hoiem's computer vision course (see below for references) and still could not explain most basic ideas in 3x more sessions.
We have used first 15 slides/lectures and Derek explained all these materials wonderfully in about 1 hour sessions. Meanwhile, Dr. Soryani, managed to ruin all those beautiful lectures by copy pasting (still failed - he just used pen for last few session to derive math equations meanwhile all ideas were copy pasted from Derek's notes as the instructor himself admitted) and using 2 hour session twice a week.
Also, TA's assignment, all have been stolen from different universities without referencing them. TA copy pasted questions without even thinking about them or even providing any solution/office hours for questions and shared assignments by taking screen shots of the assignments which was unbelievingly unreadable and as usual, I was the only person in the entire fucking class that was trying to explain to these morons that you are wrong. This managed to the situation were things got ugly and intentionally I got grades less 60% meanwhile I was sure about my answers as I HAD THE SOLUTIONS (assignments were stolen as I mentioned before). My final mark 14/20. Thanks God! I did not sell my honor!
Furthermore, I have a lot of evidences about scientific misconduct, zero management skills, zero skilled TAs, etc during this course.
PS. I did not want to share these content on github as computer course as it might confuse people about the importance or quality of materials, but I thought maybe it helps someone by saving 1 minute of his/her life. Good luck everyone!
All contents of these course have been stolen! from Derek Hoiem's CS543-Spring 2017 Computer Vision course at University of Illinois. Videos are also available but I highly suggest you to use Derek's lectures and home works which all are available at the given urls.
- Add citation
- create zenodo page