Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MapAI: Precision in Building Segmentation

Sander Riisøen Jyhne, Morten Goodwin, Per Arne Andersen, Ivar Oveland, Alexander Salveson Nossum, Karianne Ormseth, Mathilde Ørstavik, Andrew C. Flatman

TL;DR

MapAI proposes a dual-task competition for precise building segmentation from remote-sensing data, with Task 1 using aerial imagery only and Task 2 using LiDAR (with or without aerial data). The evaluation framework combines Intersection-over-Union (IoU) and Boundary IoU (BIoU) to emphasize overall accuracy and boundary precision, with per-task scores averaged to form the final ranking. The dataset includes Denmark-based training data and Norway-based test data, noting DSM and DTM differences that affect ground-truth schemas, and employs a GitHub-driven submission workflow for automated testing. The initiative targets improved generalization to small buildings across diverse environments and aims to engage students and practitioners in advancing remote-sensing segmentation.

Abstract

MapAI: Precision in Building Segmentation is a competition arranged with the Norwegian Artificial Intelligence Research Consortium (NORA) in collaboration with Centre for Artificial Intelligence Research at the University of Agder (CAIR), the Norwegian Mapping Authority, AI:Hub, Norkart, and the Danish Agency for Data Supply and Infrastructure. The competition will be held in the fall of 2022. It will be concluded at the Northern Lights Deep Learning conference focusing on the segmentation of buildings using aerial images and laser data. We propose two different tasks to segment buildings, where the first task can only utilize aerial images, while the second must use laser data (LiDAR) with or without aerial images. Furthermore, we use IoU and Boundary IoU to properly evaluate the precision of the models, with the latter being an IoU measure that evaluates the results' boundaries. We provide the participants with a training dataset and keep a test dataset for evaluation.

MapAI: Precision in Building Segmentation

TL;DR

MapAI proposes a dual-task competition for precise building segmentation from remote-sensing data, with Task 1 using aerial imagery only and Task 2 using LiDAR (with or without aerial data). The evaluation framework combines Intersection-over-Union (IoU) and Boundary IoU (BIoU) to emphasize overall accuracy and boundary precision, with per-task scores averaged to form the final ranking. The dataset includes Denmark-based training data and Norway-based test data, noting DSM and DTM differences that affect ground-truth schemas, and employs a GitHub-driven submission workflow for automated testing. The initiative targets improved generalization to small buildings across diverse environments and aims to engage students and practitioners in advancing remote-sensing segmentation.

Abstract

MapAI: Precision in Building Segmentation is a competition arranged with the Norwegian Artificial Intelligence Research Consortium (NORA) in collaboration with Centre for Artificial Intelligence Research at the University of Agder (CAIR), the Norwegian Mapping Authority, AI:Hub, Norkart, and the Danish Agency for Data Supply and Infrastructure. The competition will be held in the fall of 2022. It will be concluded at the Northern Lights Deep Learning conference focusing on the segmentation of buildings using aerial images and laser data. We propose two different tasks to segment buildings, where the first task can only utilize aerial images, while the second must use laser data (LiDAR) with or without aerial images. Furthermore, we use IoU and Boundary IoU to properly evaluate the precision of the models, with the latter being an IoU measure that evaluates the results' boundaries. We provide the participants with a training dataset and keep a test dataset for evaluation.
Paper Structure (8 sections, 2 equations, 4 figures)

This paper contains 8 sections, 2 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Dataset Details
  3. Task Descriptions
  4. Task 1: Aerial Image Segmentation Task
  5. Task 2: Laser Data Segmentation Task
  6. Submission
  7. Evaluation Methodology
  8. Summary

Figures (4)

  • Figure 1: Aerial image sample from the training dataset
  • Figure 2: Lidar sample from the training dataset
  • Figure 3: Mask sample from the training dataset
  • Figure 4: The Boundary Intersection-over-Union (BIoU) used to measure the accuracy of the segmentation boundary biou