Dynamic Day Night Anime Wallpaper Tool
What is it?
With Core ML
How to use?
- You can compile and run the sources directly.
- Or you can download releases directly
- When running: Select a folder that contains only the images (with no sub-directories) -> Wait for ML to recognize -> Correct the wrong recognitions by clicking on the image name -> Minimize the program (do not close the window)
What does the Machine Learning Module recognize?
It's trained with a limited set of features in the anime wallpaper, like the sky, the cloud, and nearby environment to help the machine differentiate between daytime and night.
What can be improved? (TO-DO)
- PR is welcomed! ;-)
- The ML module was only trained with a limited amount of images. Could add more images. Also, it's possible to train additional images to help it recognize other types of images.
- The app should be able to run in the background with a icon and menu on the system bar.
- The UI interface needs more work.
Presents the user basic interface and file selector to allow users to choose a path. After a path is chosen, it will proform a ML Vision request for the images, and show a progress bar.
This view shows the result of the ML Vision recognition through a TableView. You can click on the cell (image name) to switch its status (day/ night)
Why make this app?
I make a lot of other apps and have projects. This project is built just for fun and for experimenting the possible applications of Core ML.
Please refer to the LICENSE file in the root directly by clicking here. Your contribution to this repo will also be cited if, in the future, this app is published by me.