Computer Vision — Week 7
• publicTable of contents
Hey! My name’s Michel Liao. I’m a computer science first-year at Princeton University, hoping to get a PhD in computer vision. Join me in my CV journey!
GitHub: https://github.com/Michel-Liao
Videos/Lectures
I watched this image stitching series by Professor Shree Nayar. Notes here!
I’m working on this optical flow series by Prof. Nayar.
Mini Projects
I worked on a homography assignment created by Erich Liang. You can find the code here, but I won’t be implementing the rest of the assignment unless I have free time.
Course Progress
Finished Coursera’s Unsupervised Learning course, completing their ML series! Find my notes on unsupervised learning here.
Now, I’m on their deep learning series.
Paper of the Week
No paper this week.
Insights
Most of the insights for this week are in the notes I’ve linked above!
- Homogeneous coordinates allow us to represent translations in terms of a 3x3 transformation matrix.
- Reinforcement learning doesn’t teach the model exactly how it should act. Rather, it just defines certain goals that are good (rewards) and the NN will figure out how it should act.
Going Forward
I’m going to shift my focus to Prof. Nayar’s optical flow lectures and read a paper on RAFT. I’ll work on the deep learning series, too.
