According to your query, this may be helpful

Article A generalized supervised classification scheme to produce pr...

If the response is helpful then click on Recommend to help others. Thank you.

Google engine / Google Colab can run most of the state-of-the-art CNN for classification. However, training/transfer learning is a bit tricky. Then the comparison, it's simple as per your question, use nay image dataset e.g. ImageNet and compare their accuracy, PPV and Error rate, etc.

Thanks Khurram,

I am working on a regional scale with a classification of land use using GEE: Crops (Fruit trees, vegetables, pastures, etc.), natural field (forests, shrubs, grasslands), etc. These coverages have very strong seasonal variations and the idea is to select a supervised classification method that can capture these variations and improve the accuracy of the classification.

According to your query, this may be helpful

Article A generalized supervised classification scheme to produce pr...

If the response is helpful then click on Recommend to help others. Thank you.