https://t.co/oOcs3ddmXU helped me find a bug in this code in under 5 minutes of looking through the repo and got it merged by Phil Wang (aka lucidrains): 🤯https://t.co/IPAvrv2stA
BWlab announces release of the Mid-Level Vision Toolbox: https://t.co/PgEYflM4Hn
- line drawings from photos
- contour and junction properties
- medial axis-based local parallelism, mirror symmetry, contour separation
- visualization tools
- data sets for TorontoScenes and IAPS
What makes Morteza et al.'s work so cool is that it reproduces human-like contour completion in simple shapes and also allows for contour-based filling-in in real-world images. Check it out! The arXiv version is now available at: https://t.co/2XOgFpQZkq
Hi all, I recently made a video about machine learning fairness, and how we can mitigate biases in data using adversarial learning! https://t.co/xEKz0RqWRb
@ProbablyLearn
Had a great time at Twitter's #sfsummervibes open house event! Having a diverse team drives innovation, as everyone will have a unique way of looking at things. Glad to meet Twitter employees from all sorts of backgrounds -- each bringing a unique personal flare to the table!