iSAID: A Large-scale Dataset for Instance Segmentation in Aerial Images

TL;DR

The paper introduces iSAID, a large-scale, densely annotated dataset for instance segmentation in aerial images, addressing unique challenges like numerous tiny objects and large-scale variations, and benchmarks existing methods on this new dataset.

Contribution

The creation of the first comprehensive dataset for instance segmentation in aerial imagery, with extensive annotations and a benchmark for evaluating segmentation methods.

Findings

Existing methods perform suboptimally on aerial images.

iSAID contains 15 times more categories than previous datasets.

Specialized solutions are needed for aerial image instance segmentation.

Abstract

Existing Earth Vision datasets are either suitable for semantic segmentation or object detection. In this work, we introduce the first benchmark dataset for instance segmentation in aerial imagery that combines instance-level object detection and pixel-level segmentation tasks. In comparison to instance segmentation in natural scenes, aerial images present unique challenges e.g., a huge number of instances per image, large object-scale variations and abundant tiny objects. Our large-scale and densely annotated Instance Segmentation in Aerial Images Dataset (iSAID) comes with 655,451 object instances for 15 categories across 2,806 high-resolution images. Such precise per-pixel annotations for each instance ensure accurate localization that is essential for detailed scene analysis. Compared to existing small-scale aerial image based instance segmentation datasets, iSAID contains 15$\times$ the number of object categories and 5$\times$ the number of instances. We benchmark our dataset using two popular instance segmentation approaches for natural images, namely Mask R-CNN and PANet. In our experiments we show that direct application of off-the-shelf Mask R-CNN and PANet on aerial images provide suboptimal instance segmentation results, thus requiring specialized solutions from the research community. The dataset is publicly available at: https://captain-whu.github.io/iSAID/index.html