DFAN: Dual-Branch Feature Alignment Network for Domain Adaptation on Point Clouds

Shi, Liangwei; Yuan, Zhimin; Cheng, Ming; Chen, Yiping; Wang, Cheng

doi:10.1109/tgrs.2022.3171038

Cited by 7 publications

(3 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As in Ref., 66 this work utilizes a novel (gradientweighted) adversarial learning scheme. Finally, in Ref., 81 researchers utilize SAR-specific algorithms to extract key-point features and use a GNN to process these. They then fuse these features with features extracted from an autoencoder to build a classifier.…”

Section: Related Workmentioning

confidence: 99%

Graph pretraining approach to utilize synthetic data for SAR ATR

Parks,

Gauch,

Scherreik

et al. 2024

Algorithms for Synthetic Aperture Radar Imagery XXXI

View full text Add to dashboard Cite

Because of limitations in availability of synthetic aperture radar (SAR) training data, automatic target recognition (ATR) researchers have turned to the use of synthetic SAR images. Unfortunately, training neural network classifiers on this synthetic data does not yield robust models. Assuming access to limited measured SAR data, we evaluate two natural, transfer-learning approaches to solve this problem, showing that both do not successfully lead to solutions. Motivated by the successes of contrastive, representation, and metric learning, we propose a novel graph-based pretraining approach to transfer knowledge from synthetic samples to real-world scenarios. We show that this approach is applicable to three different neural network architectures obtaining improvements over the baseline approach of 19.21%, 28.70%, and 8.27% respectively. We also demonstrate that our method is robust to the choice of hyperparameters.

show abstract

Section: Related Workmentioning

confidence: 99%

Graph pretraining approach to utilize synthetic data for SAR ATR

Parks,

Gauch,

Scherreik

et al. 2024

Algorithms for Synthetic Aperture Radar Imagery XXXI

View full text Add to dashboard Cite

show abstract

“…These networks have attained discriminative feature representations, adeptly tackling the challenges inherent in point cloud object classification and segmentation tasks. Nevertheless, these deep learning methods require a large amount of labeled data for training to achieve good classification results [2], and annotating large amounts of data is time-consuming and expensive [3]. Due to the limited labeled point cloud datasets at present, the research community desires to build models with stronger generalization through limited annotated data.…”

Section: Introductionmentioning

confidence: 99%

“…Domain Adaptation Network for point cloud data (PointDAN) is a 3D point cloud unsupervised domain adaptive method [10]. This method dynamically collects local and global structures to extract more detailed features of point clouds between the source and target domains and jointly aligns the global and local features at multiple levels [2] [11].…”

Section: Introductionmentioning

confidence: 99%

PointADAM: Unsupervised Adversarial Domain Adaptation on Point Clouds With Metric Learning via Compact Feature Representation

Lu,

Yap,

Khor

2024

IEEE Access

View full text Add to dashboard Cite

Domain adaptation can mitigate the problem of limited labels in deep learning training. Nevertheless, extending the 2D domain adaptation method directly to 3D encounters challenges unique to point clouds, frequently resulting in inadequate feature alignment and a lack of discriminative features for decision boundaries. In light of this, we propose an unsupervised adversarial domain adaptation with metric learning (PointADAM) via compact feature representation. PointADAM is a two-stage architecture. In the first stage, it extracts high-level semantic and low-level local features from source and target domains based on the geometric relations of point cloud neighbors. Subsequently, it aligns the features from the two domains in an unsupervised adversarial manner, thereby obtaining a shared common feature space. In the second stage, a self-training method is employed to label the target domain with pseudo-labels. It is followed by selecting target domain samples with high confidence. To enhance discriminative feature learning, joint supervision is employed through both cross-entropy loss and center loss. Cross-entropy is utilized to calculate the probability of each category for classification and center loss is utilized to compact intra-class distribution and disperse inter-class distance, thereby improving classification accuracy by optimizing the decision boundary. Experimental results on the PointDA-10 dataset illustrate that our method performs relatively better than the state-of-the-art methods for domain adaptation on the point cloud by 3.1% in accuracy.

show abstract