Building detection from the SpaceNet dataset by using Mask RCNN.

The P3 Dataset: Pixels, Points and Polygons for Multimodal Building Vectorization

Building detection from the SpaceNet dataset using UNet.

Deep Learning Based Building Detection with Satellite Imagery

Building detection model with YOLOv10 on UAVOD-10 dataset

This initiative leverages cutting-edge machine learning technique such as Mask R-CNN to automate the identification of buildings in satellite images after disasters. Employing high-resolution Maxar imagery, our models efficiently and accurately pinpoint affected structures, enhancing the speed and effectiveness of emergency responses.

A deep learning project utilizing Mask R-CNN for building instance segmentation, openings detection, and building type classification.

Group Project for USTH Deep Learning Course 2023, using Faster R-CNN with pretrained model weight (gone wrong)

An AI-powered web application for detecting and extracting building footprints from satellite imagery.

This repository provides an implementation of semantic segmentation for road networks using PyTorch and the U-Net architecture. It focuses specifically on processing aerial images from the Massachusetts dataset.

Sistema automatizado de extracción de huellas de edificios afectados por desastres naturales mediante deep learning y análisis geoespacial. Caso de estudio: erupción volcánica de La Palma (2021). Integra segmentación de imágenes satelitales, regularización geométrica y validación SIG para respuesta ante emergencias.

Preprocessed and split version of the SSBH remote sensing dataset for building height estimation, including RGB composites, height maps, and building masks with train/valid/test manifests and ready-to-train scripts.