Visualize different ML and image processing memes
You can currently do the following:
- Generate covariance data
- Do PCA on data and visualize the eigenvectors pointing in the two direction with most variance
- Generate gaussian clusters
- Do K-Means(or K-Means++) Clustering and choose between L1, L2 and IOU as distance metrics.
- Load images and do different kinds of thresholding(binary, b/w, r/g/b).
- Find connected components in an image, based on its colors.
- Apply predefined convolution filters or customize your own filters.
- Visualize how the margins an SVM creates using either a gaussian or linear kernel changes when you add points and/or adjust an SVM's parameters.
- Generate linear data and perform linear or logistic regression on it
You need CMake and SDL2 to build this project. When you have those dependencies, do the following:
git clone [email protected]:marvrez/meme-visualizer.git
cd meme-visualizer/
mkdir build
cd build
cmake ..
makeIf the project is sucessfully built, you can run the program by using the following command:
./visualizerTODO:
- (H)DBSCAN