Goody or Baddy?
This Dropzone.js sends the image to an Azure Function that runs it through an ML.NET image classification machine learning model I trained with datasets of good and bad cartoon characters. It is fairly accurate but requires further training.
Follow me on Twitter (@LeeEnglestone) for more information.
Some example cartoon character images to download and try.
![Goody 1](images/good1.jpg)
![Goody 2](images/good2.jpg)
![Goody 3](images/good3.jpg)
![Goody 4](images/good4.jpg)
![Goody 5](images/good5.jpg)
![Goody 6](images/good6.jpg)
![Goody 7](images/good7.jpg)
![Goody 8](images/good8.jpg)
![Baddy 1](images/bad1.jpg)
![Baddy 2](images/bad2.jpg)
![Baddy 3](images/bad3.jpg)
![Baddy 4](images/bad4.jpg)
![Baddy 5](images/bad5.jpg)
![Baddy 6](images/bad6.jpg)
![Baddy 7](images/bad7.jpg)
![Baddy 8](images/bad8.jpg)
The code is available on GitHub at https://github.com/leeenglestone/GoodyOrBaddy