Unsupervised Domain Adaptation (UDA) aims to transfer knowledge acquired from the labeled source domain to the unlabeled target domain. However, the quality of samples can vary greatly. While partial samples are dominated by highquality domain-invariant class-related information, others may only contain irrelevant domain-specific information or useless random noise. Treating all samples equally may lead to negative transfer, significantly impairing the performance. To address the issue of varying sample quality, we propose an attention module to emphasize the samples that are most suitable for transfer. Within the attention module, we have designed a fuzzy inference system to assess the quality of data based on its class and domain information. Such a fuzzy inference attention module (FIA) demonstrates strong interpretability due to its consideration of the fuzzy nature inherent in class and domain information within the data. FIA also has high flexibility and extensibility as the rule base can be easily adjusted by expert knowledge. More importantly, FIA does not use any parameters requiring training and has a low overhead. This makes it fast and applicable to most existing unsupervised domain adaptation methods. The experiments on several benchmark datasets prove that FIA can bring significant improvement to existing methods

Aligning and balancing the marginal and conditional feature distributions are two critical procedures for unsupervised domain adaptation (UDA) problems. However, existing methods usually consider the former while ignoring the latter. To improve the problems of instability and imbalance, we propose the Adaptative Joint Distribution Adaptation Network (AJDAN) by analyzing the multi-modal interactions between the two types of distributions and adding a self-learning network to simultaneously balance them. Furthermore, we give higher weights to samples that are far away from the domain boundary (easy-to-classify samples) using Strong Binary Cross-Entropy (SBCE). The strong alignment strategy is adjustable and allows the network to better train easy-to-classify samples than traditional Binary Cross-Entropy (BCE) in various scenarios. The experiment shows that AJDAN with SBCE (AJDAN+S) provides an average of 68.3% accuracy on the Office-Home dataset, and 89.1% accuracy on the Office31 dataset, showing its superiority by 2~3 percent above the existing state-of-the-art methods.

Recently, single-source domain generalization (SDG) has gained popularity in medical image segmentation. As a prominent technique, adversarial image augmentation technique can generate synthetic training data that are challenging for the segmentation model to recognize. To avoid the over-augmentation problem, existing adversarial-based works often employ augmenters with relatively simple structures for medical images, typically operating at the image level, limiting the diversity of the augmented images. In this paper, we propose a Teacher-Student Instance-level Adversarial Augmentation (TSIAA) model for generalized medical image segmentation. The objective of TSIAA is to derive domain-generalizable representations by exploring out-of-source data distributions. First, we construct an Instance-level Image Augmenter (IIAG) using several Instance-level Augmentation Modules (IAMs), which are based on the learnable constrained Bézier transformation function. Compared to image-level adversarial augmentation, instance-level adversarial augmentation breaks the uniformity of augmentation rules across different structures within an image, thereby providing greater diversity. Then, TSIAA conducts Teacher-Student (TS) learning through an adversarial approach, alternating novel image augmentation and generalized representation learning. The former delves into out-of-source and plausible data, while the latter continuously updates both the student and teacher to ensure the original and augmented features maintain consistent and generalized characteristics. By integrating both strategies, our proposed TSIAA model achieves significant improvements over state-of-the-art methods in four challenging SDG tasks. The code can be accessed at https://github.com/Wangzs0228/TSIAA.

Conventional models for hyperspectral image (HSI) classification usually demand a substantial quantity of labeled training data. However, when labeled training HSIs are unavailable or have different distributions from test HSIs, many classification models tend to exhibit significant performance declines. For the cross-scene HSI classification task, unsupervised domain adaptation (UDA) technique has been widely developed. Existing discrepancy-based or adversarial-based UDA methods may fail to learn discriminative class boundaries when large class distribution shift (CDS) exists. To alleviate this limitation, we propose the model named Spatially Enhanced Refined Classifier (SERC), which includes a coarse classifier and a refined classifier. The refined classifier constructs a memory module to fuse global-spatial and spectral information simultaneously, and employs the neighborhood aggregation technique to generate the refined predictions. The refined predictions are then transferred to pseudo-labels to train the coarse classifier. Thereby, a mutually reinforcing relationship between the two classifiers is established. Furthermore, we propose the Class Distribution Match (CDM) strategy to further alleviate the serious CDS problem. Notably, SERC doesn’t require additional training parameters, which are commonly employed in existing UDA methods. Despite its simplicity, SERC achieves outstanding results. Our method has been extensively evaluated on three public HSI datasets and has shown superior performance compared to state-of-the-art (SOTA) approaches. The source code can be found at https://github.com/Wangzs0228/SERC.