No abstract

A domain adaptation method for urban scene segmentation is proposed in this work. We develop a fully convolutional tri-branch network, where two branches assign pseudo labels to images in the unlabeled target domain while the third branch is trained with supervision based on images in the pseudo-labeled target domain. The re-labeling and re-training processes alternate. With this design, the tri-branch network learns target-specific discriminative representations progressively and, as a result, the cross-domain capability of the segmenter improves. We evaluate the proposed network on largescale domain adaptation experiments using both synthetic (GTA) and real (Cityscapes) images. It is shown that our solution achieves the state-of-the-art performance and it outperforms previous methods by a significant margin.

The doubly fed induction generator (DFIG) is the most widely applied wind turbine in practice and contains a dc-link capacitance in its back-to-back converter to deliver the generated power and to provide dc-link voltage for converters. In small-signal analysis, fluctuations in dc-link voltage can be caused by perturbations injected at the ac-terminals of DFIG system and the dc-link voltage fluctuations can affect the outputs of ac-terminals in reverse. It has been found in former publications that this dc-link dynamic behavior is able to shape the small-signal characteristics of DFIG system and influence the system stability. However, the modeling and analysis for the dc-link dynamic behavior is lacked, which makes it difficult to indicate the influencing factors of dc-link dynamics and how to suppress the negative influence of weakly-damped dc-link dynamics on the DFIG system. In this paper, the mechanism of dc-link dynamics and how dc-link dynamics affect the small-signal characteristics of DFIG system are described at first. Then, an indicator function that models the dc-link dynamic behavior is firstly defined and then obtained based on harmonic linearization method. The proposed indicator function will be applied to describe the impedance shaping effect of dc-link dynamics on DFIG system, indicate the influencing factors of dc-link dynamics and analyze the influence of weakly-damped dc-link dynamic behavior. Several experiments based on Control-hardware-in-loop (CHIL) platform will also be carried out to verify the analytical models and theoretical analysis in this paper.INDEX TERMS Control-hardware-in-loop (CHIL), dc-link dynamics, doubly fed induction generator (DFIG), impedance modeling, indicator.

In recent years, with the development of deep learning, semantic segmentation for remote sensing images has gradually become a hot issue in computer vision. However, segmentation for multicategory targets is still a difficult problem. To address the issues regarding poor precision and multiple scales in different categories, we propose a UNet, based on multi-attention (MA-UNet). Specifically, we propose a residual encoder, based on a simple attention module, to improve the extraction capability of the backbone for fine-grained features. By using multi-head self-attention for the lowest level feature, the semantic representation of the given feature map is reconstructed, further implementing fine-grained segmentation for different categories of pixels. Then, to address the problem of multiple scales in different categories, we increase the number of down-sampling to subdivide the feature sizes of the target at different scales, and use channel attention and spatial attention in different feature fusion stages, to better fuse the feature information of the target at different scales. We conducted experiments on the WHDLD datasets and DLRSD datasets. The results show that, with multiple visual attention feature enhancements, our method achieves 63.94% mean intersection over union (IOU) on the WHDLD datasets; this result is 4.27% higher than that of UNet, and on the DLRSD datasets, the mean IOU of our methods improves UNet’s 56.17% to 61.90%, while exceeding those of other advanced methods. The implementation code is available on the following Github Link.

Deep learning approaches based on convolutional neural networks (CNNs) have been successful in solving a number of problems in medical imaging, including image segmentation. In recent years, it has been shown that CNNs are vulnerable to attacks in which the input image is perturbed by relatively small amounts of noise so that the CNN is no longer able to perform a segmentation of the perturbed image with sufficient accuracy. Therefore, exploring methods on how to attack CNN-based models as well as how to defend models against attacks have become a popular topic as this also provides insights into the performance and generalization abilities of CNNs. However, most of the existing work assumes unrealistic attack models, i.e. the resulting attacks were specified in advance. In this paper, we propose a novel approach for generating adversarial examples to attack CNN-based segmentation models for medical images. Our approach has three key features: 1) The generated adversarial examples exhibit anatomical variations (in form of deformations) as well as appearance perturbations; 2) The adversarial examples attack segmentation models so that the Dice scores decrease by a pre-specified amount; 3) The attack is not required to be specified beforehand. We have evaluated our approach on CNN-based approaches for the multi-organ segmentation problem in 2D CT images. We show that the proposed approach can be used to attack different CNN-based segmentation models.