Training a single multi-class convolutional segmentation network using multiple datasets with heterogeneous labels: Preliminary results

Fanjie Kong, Cheng Chen, Bohao Huang, Leslie M. Collins, Kyle Bradbury, Jordan M. Malof

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Segmentation convolutional neural networks (CNNs) are now popular for the semantic segmentation (i.e., dense pixel-wise labeling) of remote sensing imagery, such as color or hyperspectral satellite imagery. In recent years a large number of hand-labeled datasets of overhead imagery have emerged, leading to breakthrough performance for CNNs. However, these datasets are typically used in isolation of one another because they are either (i) annotated with heterogeneous object type labels, or (ii) they are collected over different geographic areas. This imposes a major bottleneck on the value of these datasets. In this work we present what we call a class-asymmetric loss function that makes it possible to train a single multi-class network using multiple datasets that are heterogeneously-labeled. We show, for example, that it is possible to train a segmentation algorithm for Buildings, roads, and background using two datasets: one annotated with buildings and one annotated with buildings. We propose a class asymmetric loss that under certain common conditions, allows for one to train models on datasets in which the target class is unlabeled.

Original languageEnglish
Title of host publication2019 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3903-3906
Number of pages4
ISBN (Electronic)9781538691540
DOIs
StatePublished - Jul 2019
Event39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019 - Yokohama, Japan
Duration: Jul 28 2019Aug 2 2019

Publication series

NameInternational Geoscience and Remote Sensing Symposium (IGARSS)

Conference

Conference39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019
Country/TerritoryJapan
CityYokohama
Period07/28/1908/2/19

Keywords

  • aerial imagery
  • building detection
  • convolutional neural networks
  • deep learning
  • semantic segmentation

Fingerprint

Dive into the research topics of 'Training a single multi-class convolutional segmentation network using multiple datasets with heterogeneous labels: Preliminary results'. Together they form a unique fingerprint.

Cite this