Apppublic

devkota/semanticSegmentationDemo

sourceHugging Faceupdated 3mo agoView on Hugging Face

0likes

App README

🌍 LandCoverAI Semantic Segmentation

This demo performs semantic segmentation on satellite imagery using a U-Net model trained on LandCover.ai dataset.

Features

—5 Land Cover Classes: Background, Buildings, Woodlands, Water, Roads
—Real-time Segmentation: Upload any satellite image
—Visual Comparison: Side-by-side view of input and segmentation mask

Model Details

—Architecture: U-Net with Batch Normalization
—Input Size: 256x256 pixels
—Training: 15 epochs on LandCover.ai dataset

Use Cases

—Urban planning
—Environmental monitoring
—Agriculture analysis
—Land use classification

Attribution

—Dataset: LandCover.ai
—Framework: PyTorch & Gradio

Created with ❤️ using Hugging Face Spaces