Robust Deep Neural Networks for Road Extraction From Remote Sensing Images

Li, Panle; He, Xiaohui; Qiao, Mengjia; Cheng, Xijie; Li, Zhiqiang; Luo, Haotian; Song, Dingjun; Li, Daidong; Hu, Shaokai; Li, Runchuan; Han, Pu; Qiu, Fangbing; Guo, Hengliang; Shang, Jiandong; Tian, Zengshan

doi:10.1109/tgrs.2020.3023112

Cited by 21 publications

(8 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Inspired by LNL methods for classification, [61] and [62] appended a probabilistic model to ordinary deep neural networks in order to capture the relationship between noisy labels and their latent true counterparts for road and building extraction. Nevertheless, more common and effective solutions employ robust loss functions such as bootstrapping [63], consistency constraints [11], [64], and loss reweighting with weights of each sample estimated by an attention mechanism [65] or reliability [16].…”

Section: B Lnl For Image Segmentation Tasksmentioning

confidence: 99%

AIO2: Online Correction of Object Labels for Deep Learning With Incomplete Annotation in Remote Sensing Image Segmentation

Liu,

Albrecht,

Wang

et al. 2024

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

While the volume of remote sensing data is increasing daily, deep learning in Earth Observation faces lack of accurate annotations for supervised optimization. Crowdsourcing projects such as OpenStreetMap distribute the annotation load to their community. However, such annotation inevitably generates noise due to insufficient control of the label quality, lack of annotators, frequent changes of the Earth's surface as a result of natural disasters and urban development, among many other factors. We present Adaptively trIggered Online Object-wise correction (AIO2) to address annotation noise induced by incomplete label sets. AIO2 features an Adaptive Correction Trigger (ACT) module that avoids label correction when the model training under-or overfits, and an Online Object-wise Correction (O2C) methodology that employs spatial information for automated label modification. AIO2 utilizes a mean teacher model to enhance training robustness with noisy labels to both stabilize the training accuracy curve for fitting in ACT and provide pseudo labels for correction in O2C. Moreover, O2C is implemented online without the need to store updated labels every training epoch. We validate our approach on two building footprint segmentation datasets with different spatial resolutions. Experimental results with varying degrees of building label noise demonstrate the robustness of AIO2. Source code will be available at https: //github.com/zhu-xlab/AIO2.git.

show abstract

Section: B Lnl For Image Segmentation Tasksmentioning

confidence: 99%

AIO2: Online Correction of Object Labels for Deep Learning With Incomplete Annotation in Remote Sensing Image Segmentation

Liu,

Albrecht,

Wang

et al. 2024

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Approaches dealing with Data Noise: NSCT [95], HRO [97], PIQE [98], ACL-CNN [99], HS2P [9]. Approaches dealing with Label Noise: tRNSL [102], NTDNE [103], AF2GNN [105], RSSC-ETDL [106], CSHLC [107], I-FPFN-EM [108], FFCTL [109], RS-COCL-NLF [110], RVgg19 [111]. Table II provides more information about datasets, modalities and backbone of each reviewed methods.…”

Section: A Data Noisementioning

confidence: 99%

“…Nevertheless, a clean set of true labels calibrated by the author is required for the label correction learning process, which is costly to apply to other large-scale datasets. A promising research direction is the combination of noise minimization and correction, as explored in recent studies [109]- [111]. Most recent approaches commonly involve modeling uncertainty using probabilistic methods to address label noise [124].…”

Section: B Label Noisementioning

confidence: 99%

Multimodal Colearning Meets Remote Sensing: Taxonomy, State of the Art, and Future Works

Kieu,

Nguyen,

Nazib

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

In remote sensing (RS), multiple modalities of data are usually available, e.g., RGB, Multispectral, Hyperspectral, LiDAR, and SAR. Multimodal machine learning systems, which fuse these rich multimodal data modalities, have shown better performance compared to unimodal systems. Most multimodal research assumes that all modalities are present, aligned, and noiseless during training and testing time. However, in real-world scenarios, it is common to observe that one or more modalities are missing, noisy, and non-aligned, in either training or testing or both. In addition, acquiring largescale, noise-free annotations is expensive, as a result, lacking sufficient annotated datasets or having to deal with inconsistent labels are open challenges. These challenges can be addressed under a learning paradigm called multimodal co-learning.This paper focuses on multimodal co-learning techniques for remote sensing data. We first review what data modalities are available in the remote sensing domain and the key benefits and challenges of combining multimodal data in the remote sensing context. We then review the remote sensing tasks that would benefit from multimodal processing including classification, segmentation, target detection, anomaly detection, and temporal change detection. We then dive deeper into technical details by reviewing more than 200 recent efforts in this area and provide a comprehensive taxonomy to systematically review state-of-the-art approaches in 4 key co-learning challenges including missing modalities, noisy modalities, limited modality annotations, and weakly-paired modalities. Based on these insights, we propose emerging research directions to inform potential future research in multimodal co-learning for remote sensing.

show abstract

“…[21] incorporated noiserobust loss functions into the CNN architecture to reduce the effect of omission and registration errors in OSM. Inspired by [44], [45] and [46] adopted an extra noise adaptation layer to model a correlation between clean and noisy labels. They successfully improved the performance of the baseline model's performance with the supervision of noisy OSM labels.…”

Section: B Noisy Supervisionmentioning

confidence: 99%

Self-Filtered Learning for Semantic Segmentation of Buildings in Remote Sensing Imagery With Noisy Labels

Song

Yang

Jung

2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Not all building labels for training improve the performance of the deep learning model. Some labels can be falsely labeled or too ambiguous to represent their groundtruths, resulting in poor performance of the model. For example, building labels in OpenStreetMap (OSM) and Microsoft Building Footprints (MBF) are publicly available training sources that have great potential to train deep models, but directly using those labels for training can limit the model's performance as their labels are incomplete and inaccurate, called noisy labels. This paper presents self-filtered learning (SFL) that helps a deep model learn well with noisy labels for building extraction in remote sensing images. SFL iteratively filters out noisy labels during the training process based on loss of samples. Through a multi-round manner, SFL makes a deep model learn progressively more on refined samples from which the noisy labels have been removed. Extensive experiments with the simulated noisy map as well as real-world noisy maps, OSM and MBF, showed that SFL can improve the deep model's performance in diverse error types and different noise levels.

show abstract

Robust Deep Neural Networks for Road Extraction From Remote Sensing Images

Cited by 21 publications

References 39 publications

AIO2: Online Correction of Object Labels for Deep Learning With Incomplete Annotation in Remote Sensing Image Segmentation

AIO2: Online Correction of Object Labels for Deep Learning With Incomplete Annotation in Remote Sensing Image Segmentation

Multimodal Colearning Meets Remote Sensing: Taxonomy, State of the Art, and Future Works

Self-Filtered Learning for Semantic Segmentation of Buildings in Remote Sensing Imagery With Noisy Labels

Contact Info

Product

Resources

About