Object detection typically assumes that training and test data are drawn from an identical distribution, which, however, does not always hold in practice. Such a distribution mismatch will lead to a significant performance drop. In this work, we aim to improve the cross-domain robustness of object detection. We tackle the domain shift on two levels: 1) the image-level shift, such as image style, illumination, etc., and 2) the instance-level shift, such as object appearance, size, etc. We build our approach based on the recent state-of-the-art Faster R-CNN model, and design two domain adaptation components, on image level and instance level, to reduce the domain discrepancy. The two domain adaptation components are based on H-divergence theory, and are implemented by learning a domain classifier in adversarial training manner. The domain classifiers on different levels are further reinforced with a consistency regularization to learn a domain-invariant region proposal network (RPN) in the Faster R-CNN model. We evaluate our newly proposed approach using multiple datasets including Cityscapes, KITTI, SIM10K, etc. The results demonstrate the effectiveness of our proposed approach for robust object detection in various domain shift scenarios.

BACKGROUND: On January 20, 2020, a new coronavirus epidemic with human-to-human transmission was officially declared by the Chinese government, which caused significant public panic in China. In light of the coronavirus disease 2019 outbreak, pregnant women may be particularly vulnerable and in special need for preventive mental health strategies. Thus far, no reports exist to investigate the mental health response of pregnant women to the coronavirus disease 2019 outbreak. OBJECTIVE: This study aimed to examine the impact of coronavirus disease 2019 outbreak on the prevalence of depressive and anxiety symptoms and the corresponding risk factors among pregnant women across China. STUDY DESIGN: A multicenter, cross-sectional study was initiated in early December 2019 to identify mental health concerns in pregnancy using the Edinburgh Postnatal Depression Scale. This study provided a unique opportunity to compare the mental status of pregnant women before and after the declaration of the coronavirus disease 2019 epidemic. A total of 4124 pregnant women during their third trimester from 25 hospitals in 10 provinces across China were examined in this crosssectional study from January 1, 2020, to February 9, 2020. Of these women, 1285 were assessed after January 20, 2020, when the coronavirus epidemic was publicly declared and 2839 were assessed before this pivotal time point. The internationally recommended Edinburgh Postnatal Depression Scale was used to assess maternal depression and anxiety symptoms. Prevalence rates and risk factors were compared between the pre-and poststudy groups. RESULTS: Pregnant women assessed after the declaration of coronavirus disease 2019 epidemic had significantly higher rates of depressive symptoms (26.0% vs 29.6%, P¼.02) than women assessed before the epidemic declaration. These women were also more likely to have thoughts of self-harm (P¼.005). The depressive rates were positively associated with the number of newly confirmed cases of coronavirus disease 2019 (P¼.003), suspected infections (P¼.004), and deaths per day (P¼.001). Pregnant women who were underweight before pregnancy, primiparous, younger than 35 years, employed full time, in middle income category, and had appropriate living space were at increased risk for developing depressive and anxiety symptoms during the outbreak. CONCLUSION: Major life-threatening public health events such as the coronavirus disease 2019 outbreak may increase the risk for mental illness among pregnant women, including thoughts of self-harm. Strategies targeting maternal stress and isolation such as effective risk communication and the provision of psychological first aid may be particularly useful to prevent negative outcomes for women and their fetuses.

The early inflammatory response to spinal cord injury (SCI) causes significant secondary damage. Strategies that nonselectively suppress inflammation have not improved outcomes after SCI, perhaps because inflammation has both adverse and beneficial effects after SCI. We have shown that the selective, time-limited action of a monoclonal antibody (mAb) to the CD11d subunit of the CD11d/CD18 integrin, delivered intravenously during the first 48 hr after SCI in rats, markedly decreases the infiltration of neutrophils and delays the entry of hematogenous monocyte-macrophages into the injured cord. We hypothesized that this targeted strategy would lead to neuroprotection and improved neurological outcomes. In this study the development of chronic pain was detected in rats by assessing mechanical allodynia on the trunk and hindpaws 2 weeks to 3 months after a clinically relevant clip-compression SCI at the twelfth thoracic segment. The anti-CD11d mAb treatment reduced this pain by half. Motor performance also improved as rats were able to plantar-place their hindpaws and use them for weight support instead of sweeping movements only. Improved cardiovascular outcome was shown after SCI at the fourth thoracic segment by significant decreases in autonomic dysreflexia. Locomotor performance was also improved. These functional changes correlated with significantly greater amounts and increased organization of myelin and neurofilament near the lesion. The improved neurological recovery after the specific reduction of early inflammation after SCI demonstrates that this selective strategy increases tissue at the injury site and improves its functional capacity. This early neuroprotective treatment would be an ideal foundation for building later cell-based therapies.

This paper tackles the problem of video object segmentation, given some user annotation which indicates the object of interest. The problem is formulated as pixel-wise retrieval in a learned embedding space: we embed pixels of the same object instance into the vicinity of each other, using a fully convolutional network trained by a modified triplet loss as the embedding model. Then the annotated pixels are set as reference and the rest of the pixels are classified using a nearest-neighbor approach. The proposed method supports different kinds of user input such as segmentation mask in the first frame (semi-supervised scenario), or a sparse set of clicked points (interactive scenario). In the semi-supervised scenario, we achieve results competitive with the state of the art but at a fraction of computation cost (275 milliseconds per frame). In the interactive scenario where the user is able to refine their input iteratively, the proposed method provides instant response to each input, and reaches comparable quality to competing methods with much less interaction.

Exploiting synthetic data to learn deep models has attracted increasing attention in recent years. However, the intrinsic domain difference between synthetic and real images usually causes a significant performance drop when applying the learned model to real world scenarios. This is mainly due to two reasons: 1) the model overfits to synthetic images, making the convolutional filters incompetent to extract informative representation for real images; 2) there is a distribution difference between synthetic and real data, which is also known as the domain adaptation problem. To this end, we propose a new reality oriented adaptation approach for urban scene semantic segmentation by learning from synthetic data. First, we propose a target guided distillation approach to learn the real image style, which is achieved by training the segmentation model to imitate a pretrained real style model using real images. Second, we further take advantage of the intrinsic spatial structure presented in urban scene images, and propose a spatialaware adaptation scheme to effectively align the distribution of two domains. These two modules can be readily integrated with existing state-of-the-art semantic segmentation networks to improve their generalizability when adapting from synthetic to real urban scenes. We evaluate the proposed method on Cityscapes dataset by adapting from GTAV and SYNTHIA datasets, where the results demonstrate the effectiveness of our method.• real distribution orientation: To deal with the dis-

High-resolution (HR) magnetic resonance images (MRI) provide detailed anatomical information important for clinical application and quantitative image analysis. However, HR MRI conventionally comes at the cost of longer scan time, smaller spatial coverage, and lower signal-to-noise ratio (SNR). Recent studies have shown that single image super-resolution (SISR), a technique to recover HR details from one single low-resolution (LR) input image, could provide high quality image details with the help of advanced deep convolutional neural networks (CNN). However, deep neural networks consume memory heavily and run slowly, especially in 3D settings. In this paper, we propose a novel 3D neural network design, namely a multi-level densely connected super-resolution network (mDCSRN) with generative adversarial network (GAN)-guided training. The mDCSRN trains and inferences quickly, and the GAN promotes realistic output hardly distinguishable from original HR images. Our results from experiments on a dataset with 1,113 subjects shows that our new architecture outperforms other popular deep learning methods in recovering 4x resolutiondowngraded images and runs 6x faster.

In this work, we present a domain flow generation(DLOW) model to bridge two different domains by generating a continuous sequence of intermediate domains flowing from one domain to the other. The benefits of our DLOW model are two-fold. First, it is able to transfer source images into different styles in the intermediate domains. The transferred images smoothly bridge the gap between source and target domains, thus easing the domain adaptation task. Second, when multiple target domains are provided for training, our DLOW model is also able to generate new styles of images that are unseen in the training data. We implement our DLOW model based on CycleGAN. A domainness variable is introduced to guide the model to generate the desired intermediate domain images. In the inference phase, a flow of various styles of images can be obtained by varying the domainness variable. We demonstrate the effectiveness of our model for both cross-domain semantic segmentation and the style generalization tasks on benchmark datasets. Our implementation is available at

We present GLNet, a self-supervised framework for learning depth, optical flow, camera pose and intrinsic parameters from monocular video -addressing the difficulty of acquiring realistic ground-truth for such tasks. We propose three contributions: 1) we design new loss functions that capture multiple geometric constraints (e.g. epipolar geometry) as well as an adaptive photometric loss that supports multiple moving objects, rigid and non-rigid, 2) we extend the model such that it predicts camera intrinsics, making it applicable to uncalibrated video, and 3) we propose several online refinement strategies that rely on the symmetry of our self-supervised loss in training and testing, in particular optimizing model parameters and/or the output of different tasks, thus leveraging their mutual interactions. The idea of jointly optimizing the system output, under all geometric and photometric constraints can be viewed as a dense generalization of classical bundle adjustment. We demonstrate the effectiveness of our method on KITTI and Cityscapes, where we outperform previous self-supervised approaches on multiple tasks. We also show good generalization for transfer learning in YouTube videos.