Procedural Synthesis of Remote Sensing Images for Robust Change Detection with Neural Networks

Segmentation of tumors in brain MRI images is a challenging task, where most recent methods demand large volumes of data with pixel-level annotations, which are generally costly to obtain. In contrast, image-level annotations, where only the presence of lesion is marked, are generally cheap, generated in far larger volumes compared to pixel-level labels, and contain less labeling noise. In the context of brain tumor segmentation, both pixel-level and image-level annotations are commonly available; thus, a natural question arises whether a segmentation procedure could take advantage of both. In the present work we: 1) propose a learning-based framework that allows simultaneous usage of both pixel-and image-level annotations in MRI images to learn a segmentation model for brain tumor; 2) study the influence of comparative amounts of pixel-and image-level annotations on the quality of brain tumor segmentation; 3) compare our approach to the traditional fully-supervised approach and show that the performance of our method in terms of segmentation quality may be competitive.

show abstract

“…A challenging task would be to test the developed capabilities on other types of image data, e.g. from remote sensing applications [25], [26].…”

Section: Discussionmentioning

confidence: 99%

Weakly Supervised Fine Tuning Approach for Brain Tumor Segmentation Problem

Павлов

Artemov

Sharaev

et al. 2019

2019 18th IEEE International Conference on Machine Learning and Applications (ICMLA)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Meanwhile, in order to embed pixel-wise correspondences, graphical model based algorithms are proposed [25], [26], which can model the spatial regularity during the optimization process. Further to expand the model capacities, deep learning networks [5], [27]- [30] are designed with elaborate structures to depict more diverse scenes by searching optimal parameters, where huge parameter space ensures stronger model capacities. Moreover, as illustrated in Sec.1, considering in the discrepancy across land-cover distributions in input images could provide extra information when depicting land-cover distributions in some cases shown in Fig.…”

Section: Location Of Changed Regionsmentioning

confidence: 99%

Asymmetric Siamese Networks for Semantic Change Detection in Aerial Images

Yang

Xia

Liu

et al. 2022

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Given two multi-temporal aerial images, semantic change detection aims to locate the land-cover variations and identify their categories with pixel-wise boundaries. The problem has demonstrated promising potentials in many earth vision related tasks, such as precise urban planning and natural resource management. Existing state-of-the-art algorithms mainly identify the changed pixels through symmetric modules, which would suffer from categorical ambiguity caused by changes related to totally different landcover distributions. In this paper, we present an asymmetric siamese network (ASN) to locate and identify semantic changes through feature pairs obtained from modules of widely different structures, which involve different spatial ranges and quantities of parameters to factor in the discrepancy across different land-cover distributions. To better train and evaluate our model, we create a large-scale well-annotated SEmantic Change detectiON Dataset (SECOND), while an adaptive threshold learning (ATL) module and a separated kappa (SeK) coefficient are proposed to alleviate the influences of label imbalance in model training and evaluation. The experimental results demonstrate that the proposed model can stably outperform the state-of-the-art algorithms with different encoder backbones.

show abstract

“…Subsequently, they fine-tune their model with a limited amount of real-world data and show a significant improvement of the results compared to a training only on real-world data. (Kolos et al, 2019) present another method to generate realistic synthetic remote sensing data: They use the Esri-City-Engine with cartographic data from OSM for geometry and the game engine Unity for data rendering. Their results for change detection, using a Siamese U-Net structure, also improve the performance and robustness of models for remote sensing applications with limited training data.…”

Section: Related Workmentioning

confidence: 99%

“…There are different strategies to avoid manual labelling: Using domain adaptation (DA) techniques a classifier that was trained on data obtained under different conditions is adjusted to the new data (Wang, Deng, 2018). Other approaches, like (Kemker, Kanan, 2017) or (Kolos et al, 2019) generate synthetic training data to pre-train the classifier. If there is no available data from a different domain weakly labeled data can be used as well.…”

Section: Introductionmentioning

confidence: 99%

Automatically Generated Training Data for Land Cover Classification With CNNS Using Sentinel-2 Images

Voelsen

Bostelmann²,

Maas³

et al. 2020

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. Pixel-wise classification of remote sensing imagery is highly interesting for tasks like land cover classification or change detection. The acquisition of large training data sets for these tasks is challenging, but necessary to obtain good results with deep learning algorithms such as convolutional neural networks (CNN). In this paper we present a method for the automatic generation of a large amount of training data by combining satellite imagery with reference data from an available geospatial database. Due to this combination of different data sources the resulting training data contain a certain amount of incorrect labels. We evaluate the influence of this so called label noise regarding the time difference between acquisition of the two data sources, the amount of training data and the class structure. We combine Sentinel-2 images with reference data from a geospatial database provided by the German Land Survey Office of Lower Saxony (LGLN). With different training sets we train a fully convolutional neural network (FCN) and classify four land cover classes (Building, Agriculture, Forest, Water). Our results show that the errors in the training samples do not have a large influence on the resulting classifiers. This is probably due to the fact that the noise is randomly distributed and thus, neighbours of incorrect samples are predominantly correct. As expected, a larger amount of training data improves the results, especially for the less well represented classes. Other influences are different illuminations conditions and seasonal effects during data acquisition. To better adapt the classifier to these different conditions they should also be included in the training data.

show abstract

Procedural Synthesis of Remote Sensing Images for Robust Change Detection with Neural Networks

Cited by 8 publications

References 58 publications

Weakly Supervised Fine Tuning Approach for Brain Tumor Segmentation Problem

Weakly Supervised Fine Tuning Approach for Brain Tumor Segmentation Problem

Asymmetric Siamese Networks for Semantic Change Detection in Aerial Images

Automatically Generated Training Data for Land Cover Classification With CNNS Using Sentinel-2 Images

Contact Info

Product

Resources

About