Maskformer Satellite Trees

🚀 Experimental Model for Satellite Image Vegetation Segmentation

This is an experimental model designed to segment vegetation in satellite images. It simply labels pixels as either "vegetation" or "other".

🚀 Quick Start

This model is created by finetuning the facebook/maskformer-swin-base-ade model. It is trained with a small number (~25) of manually labeled satellite images of urban - ish areas.

✨ Features

• Simple Labeling: Labels pixels in satellite images as either "vegetation" or "other".
• Finetuning: Based on the facebook/maskformer-swin-base-ade model.

📚 Documentation

🔍 BIAS WARNING

This model was developed for a personal art and urbanism project. The training set, while including images from geographically diverse cities of personal importance, is not exhaustive. There are no images from cities in Asia, Africa, Central America, or Oceania.

The urban areas in the training set are from or around the following cities:

• São Paulo, BR
• Rio de Janeiro, BR
• New York, US
• Pittsburgh, US
• Oakland, US
• Berlin, DE
• Milan, IT
• Riyadh, SA

🔍 EVALUATION WARNING

Anecdotally, this model seems more precise than the original facebook/maskformer-swin-base-ade model when used to obtain masks for vegetation. It can work with images from other cities, but the success criteria is qualitative.

📄 License

This project is licensed under the CC BY - NC - SA 4.0 license.