Resources for getting started with the vision team.
Getting remote repo on local pc
- cd <directory> # Change into a directory you want the folder in
- git clone <url> # Download the remote repository to local pc
Before working
- git pull # pull other peoples work to your local machine
- <edit code>
- git status # see what git sees
- git add <files you want updated on repo>
- git commit -m "Description of changes here."
- git push # push your updates to the remote repo
Numpy Basics -- Should check these out before anything else!
-
Monte carlo simulation of the game craps. Example here.
-
Image convolver with Numpy -- blurring, sobel operator for edge detection.
Interested in Autonomy?
-
Flocking algorithm, all agents think independantly but create emergent behavior.
-
Ant colony simulator, trying to mimic the use of pheremones to create intelligent, efficient communities.
Interested in writing code for the Graphics Card?
-
Follow this CUDA tutorial.
-
Write a voroni tesselation program with CUDA.
-
Compute planet orbits really fast with CUDA.
Interested in classic Computer Vision methods?
CV w/ Scipy and OpenCV Tutorial
-
Shrink images to 1/2, 2/3, 3/4 scale with selective column dropping.
-
Expand images to 2x, 3x scale with interpolation.
-
Use scipy generic_filter to simulate a cellular automaton game. Wolfram 1D Automaton, Game of Life, Langton's Ant, Wireworld.
-
Use scipy convolve to create an edge detector, blur filter or a shape detector on space images(?).
-
Perform image skeletonization or clean up images with erosion and dialation convolutions.
-
Track cells or any simple object with blob tracking, might use mean shift. This likely applicable to our Mission 9 tasks. This blog post.
Interested in Machine Learning?
-
Write a K-Nearest Neighbors model to predict the best background color to go with any given text color using this already written GUI or your own interface. Alternatively write a model for automated weather prediction.
-
Use k-means unsupervised clustering algorithm to change the color of tomatoes in a salad to blue with sklearn.
-
Create a reinforcement learning algorithm to play some openai-gym games, ie q-learning! Sample code.
-
Write a neural network from scratch or with pytorch. These 3Blue1Brown videos may help. Michael Nielsen book.