Computer Vision — Week 8
• 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 finished this optical flow series by Professor Shree Nayar. Notes here!
Mini Projects
None. Focused on paper-reading this week.
Course Progress
Finished Coursera’s Neural Networks and Deep Learning course! Notes are here.
Now I’m on Improving Deep Neural Networks: Hyperparameter Tuning, Regularization, and Optimization.
Paper of the Week
Both papers were sent to me by Erich Liang.
How to Read a Paper - Srinivasan Keshav
The Three-Pass Approach:
- First pass: 5-10 minutes
- Do not annotate
- Carefully read the title, abstract, and introduction
- Read the headings
- Read the conclusions
- Glance over references and mentally note which ones you’ve read
- Second pass: 1 hour
- Annotate
- Read the entire paper more carefully but ignore details like proofs
- Pay attention to figures
- Mark unread references for later reading
- Third pass: 4-5 hours
- Reimplement the paper
- Challenge assumptions and note ideas for future work
RAFT: Recurrent All-Pairs Field Transforms for Optical Flow - Zachary Teed and Jia Deng
I’m still working on reading and making a presentation for this paper. Presentation link is here.
Going Forward
I’m going to continue reading RAFT and working on Coursera. If I have time, I’ll implement logistic regression from scratch, post about cross-correlation from scratch, or build a NN with PyTorch.
