This project is part of the course TNM097 - Image Reproduction at LiU. It is implemented in MATLAB and transforms any input image into an image composed of Pokémon characters using k-means clustering to determine the appropriate Pokémon for each pixel. To view samples in full resolution, visit the /reproduced-images folder.
To try it out yourself, follow these steps:
- Clone this repository.
- Open it in MATLAB.
- Make sure the /matlab folder is on the MATLAB path (right-click the folder to add it to path).
- Open the Command Window.
- Call the create_poke_image function with the appropriate arguments:
create_poke_image("./path/to/image.png", "./pokemon-images/genX");
You need to provide the path to an input image and the path to a Pokémon database folder. Currently, only .png files are supported for input images. Multiple database options are available within the /pokemon-images folder. Note that more Pokémon images in the database will increase the loading time. To reduce the number of similar Pokémon images, use the reduce_db.m MATLAB function to create a new folder with fewer Pokémon. For more details, refer to the function.
