Today, I wrote about yesterday...

Table of Contents

• 1. Update
• 2. Computer Vision: Some Findings

1. Update

I found a great plug-in for Obsidian: Image Upload Toolkit. This allows one to upload any image embedded in a note to cloud directly. This will save local storage as the images are, by default, stored in the Obsidian Vault. This is not needed if one has only few images; but I set it up, nevertheless. Writing about it and uploading the below image to test the feature out:

One annoying thing about it is that I have to manually change the local link to the remote URL, and also manually delete the image from the Obsidian vault. But, oh well...

2. Computer Vision: Some Findings

For almost a year, I've been focusing on Computer Vision-related problems at work. This domain is a little challenging, yet, quite fun to learn! One big regret I have from my school days is that I didn't study Computer Vision in-depth; the course offered back then quickly filled up before I got a chance to register. It's not as if one can't learn at work, but the focus is more on applying certain techniques to particular problems. However, I wanted to learn this field theoretically and from first principles. Here are some resources I gathered that I hope I get to visit time-to-time:

1. CS231n — Deep Learning for Computer Vision: This is possibly the gold standard for a Computer Vision course that is especially focused on explaining how to apply the latest Deep Learning models. This won't make one learn Computer Vision from the ground principles though; it is focused more on explaining Deep Learning architectures and how they can be configured to apply for image and video-based problems.

   • Spring 2025 YouTube Playlist
   • Spring 2025 Syllabus
   • XCS231N as part of Stanford's Artificial Intelligence Professional Program; brochure available here.
   • CS231n as part of Stanford's Artificial Intelligence Graduate Certificate.

2. First Principles of Computer Vision: This looks like a fantastic course, and is offered by Columbia University. This is more focused on explaining Computer Vision from first principles (exactly what I was looking to find!).

   • Course Website
   • YouTube channel
   • Monographs (PDFs)

3. Computer Vision: Algorithms and Applications, 2nd edition: This is a book offered by Richard Szeliski who has 40+ years of experience in this field. He worked for Microsoft, Facebook, and Google. He is also affiliated with the University of Washington. This book offers comprehensive information from the grounds-up and is sufficiently theoretical (perhaps, more than necessary?).

   • Book website: One can download the book for free after filling basic information on a form on the website.