Can semantic labeling methods generalize to any city? the inria aerial image labeling benchmark

By Emmanuel Maggiori, Yuliya Tarabalka, Guillaume Charpiat, Pierre Alliez. 2017

811 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology, Ocean Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

In Ocean Engineering

Citations in the corpus

Albumentations: fast and flexible image augmentations (2018, methodology): “d extraction [12] and for multi-class land segmentation [13] from satellite imagery. Fig. 3 provides examples of such image transforms applied to an image from the Inria Aerial Image Labeling dataset [14]. Corresponding binary masks of buildings present in the image are also transformed in the same ”
Incremental Learning for Semantic Segmentation of Large-Scale Remote Sensing Data (2018, methodology): “Besides, for the first problem, by following a similar strategy described in [9], we test the trained models on the data collected from completely different geographic areas than the ones we use during training.”
OpenEarthMap: A Benchmark Dataset for Global High-Resolution Land Cover Mapping (2022, methodology): “[34] Emmanuel Maggiori, Yuliya Tarabalka, Guillaume Charpiat, and Pierre Alliez.”
Object Detection and Image Segmentation with Deep Learning on Earth Observation Data: A Review-Part I: Evolution and Recent Trends (2020, background): “[126] IS multiple classes plane multispectral very high U-Net, DeepLabV3+, PSPNet, LANet (patch attention module) [167], MobileNetV2(with atrous conv) + Dual path encoder + SE modules [168] DeepGlobe-Road [169] IS road network WorldView3 multispectral high D-LinkNet (ResNet-34 + U-Net with atrous de”
Understanding Deep Learning Techniques for Image Segmentation (2019, background): “ SYNTHIA: Synthetic collection of Imagery and Annotations [178] KITTI Vision Benchmark Suite [67] Berkeley Deep Drive [212] India Driving Dataset(IDD) [202] Aerial Inria Aerial Image Labeling Dataset [134] Imaging Aerial Image Segmentation Dataset [213] ISPRS Dataset collection [57] Google Open Stre”
SEN12MS - A Curated Dataset of Georeferenced Multi-Spectral Sentinel-1/2 Imagery for Deep Learning and Data Fusion (2019, background): “Inria Aerial Image Labeling Dataset 360 1,500× 1,500 Color aerial imagery 2 classes (building / not building) for 10 cities in Austria and USA (Maggiori et al., 2017)”
DeepGlobe 2018: A Challenge to Parse the Earth through Satellite Images (2018, background): “The Inria Aerial Image Labeling covered 810 km 2 area with 30cm resolution in various European and American cities [35].”
RSVQA: Visual Question Answering for Remote Sensing Data (2020, positive): “More recently, a contest organized on the data set of building segmentation created in [15] has motivated the development of a number of new methods to improve results on this task [16].”
Land-cover classification with high-resolution remote sensing images using transferable deep models (2018, background): “Up to now, several land-cover datasets have been proposed in the community, and have advanced a lot deep-learning-based land-cover classification approaches [44–46].”
iSAID: A Large-scale Dataset for Instance Segmentation in Aerial Images (2019, background): “solutions on appropriate datasets are desired. To encourage the advancements in aerial imagery for earth observation, a few well-annotated datasets for object detection [16, 32] and semantic labeling [21, 9] have recently been introduced. However, they do not provide per-pixel accurate labelings for”
A review of deep learning methods for semantic segmentation of remote sensing imagery (2021, background): “Maggiori, Tarabalka, Charpiat, and Alliez (2017) extracted intermediate features from the network at various resolutions.”
Fully Convolutional Networks for Multisource Building Extraction From an Open Aerial and Satellite Imagery Data Set (2019, neutral): “MLP provided by [28] utilized softmax for binary labeling and provide only binary labels here.”
Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning (2022, background): “The separate low/high resolution reconstructions obtain the best performance and robustness to changes in scale.”
Multi-Task Learning for Segmentation of Building Footprints with Deep Neural Networks (2017, methodology): “Our approach incorporates boundary information of buildings and improves the segmentation results while using less memory at inference time compared to [2].”
FAIR1M: A Benchmark Dataset for Fine-grained Object Recognition in High-Resolution Remote Sensing Imagery (2021, methodology): “There are also datasets dedicated to building detection, such as the Semcity toulouse dataset [43], ISPRS benchmark on urban object detection and 3D building reconstruction dataset [44], Inria Aerial Image Labeling dataset [36] and DeepGlobe 2018 dataset [9].”