This project was developed in python2 and requires the following modules:
- scipy
- numpy
- scikit-image
- scikit-learn
These can be installed using pip install <module> for each respective module.
In order to run, a set of rendered ASCII characters in the desired font must be present in the out/ directory. These can be generated in different fonts using text.py, and these are used as the training data for the nearest-neighbors and single-sample SVM approaches.
Additionally, for the regular SVM to run, the training.dat file must be present in the base directory. This can be found in training.zip, or generated using generate_data.py.
python convert.py <image name> will convert the image file provided as a command line argument into ASCII art. The output will be directly displayed to the console along with timings using each of our three algorithms.