Search

Great progress has been witnessed for weakly supervised semantic segmentation, which aims to segment objects without dense pixel annotations. Most approaches concentrate on generating high quality pseudo labels, which are then fed into a standard segmentation model as supervision. However, such a solution has one major limitation: noise of pseudo labels is inevitable, which is unsolvable for the standard segmentation model. In this paper, we propose a credible dual-expert learning (CDL) framework to mitigate the noise of pseudo labels. Specifically, we first observe that the model predictions with different optimization loss functions will have different credible regions; thus, it is possible to make self-corrections with multiple predictions. Based on this observation, we design a dual-expert structure to mine credible predictions, which are then processed by our noise correction module to update pseudo labels in an online way. Meanwhile, to handle the case that the dual-expert produces incredible predictions for the same region, we design a relationship transfer module to provide feature relationships, enabling our noise correction module to transfer predictions from the credible regions to such incredible regions. Considering the above designs, we propose a base CDL network and an extended CDL network to satisfy different requirements. Extensive experiments show that directly replacing our model with a conventional fully supervised segmentation model, the performances of various weakly supervised semantic segmentation pipelines were boosted, achieving new state-of-the-art performances on both PASCAL VOC 2012 and MS COCO with a clear margin. Code will be available at: https://github.com/zbf1991/CDL. [ABSTRACT FROM AUTHOR]

Weakly supervised semantic segmentation is receiving great attention due to its low human annotation cost. In this paper, we aim to tackle bounding box supervised semantic segmentation, i.e., training accurate semantic segmentation models using bounding box annotations as supervision. To this end, we propose affinity attention graph neural network ($A^2$ A 2 GNN). Following previous practices, we first generate pseudo semantic-aware seeds, which are then formed into semantic graphs based on our newly proposed affinity Convolutional Neural Network (CNN). Then the built graphs are input to our $A^2$ A 2 GNN, in which an affinity attention layer is designed to acquire the short- and long- distance information from soft graph edges to accurately propagate semantic labels from the confident seeds to the unlabeled pixels. However, to guarantee the precision of the seeds, we only adopt a limited number of confident pixel seed labels for $A^2$ A 2 GNN, which may lead to insufficient supervision for training. To alleviate this issue, we further introduce a new loss function and a consistency-checking mechanism to leverage the bounding box constraint, so that more reliable guidance can be included for the model optimization. Experiments show that our approach achieves new state-of-the-art performances on Pascal VOC 2012 datasets (val: 76.5 percent, test: 75.2 percent). More importantly, our approach can be readily applied to bounding box supervised instance segmentation task or other weakly supervised semantic segmentation tasks, with state-of-the-art or comparable performance among almot all weakly supervised tasks on PASCAL VOC or COCO dataset. Our source code will be available at https://github.com/zbf1991/A2GNN. [ABSTRACT FROM AUTHOR]

Uplifting mechanisms for the Tien Shan, an active intracontinental orogenic belt, have been under debate for decades, a key issue being how the convergence has been accommodated at depth. Here we investigate the Moho structure across the Central Tien Shan by common‐conversion‐point imaging and H‐k‐c stacking of receiver functions from a dense array. The observed Moho exhibits distinct characteristics among subblocks. A southward‐dipping diffuse Moho is imaged in the South‐Central Tien Shan (SCTS), in contrast with the relatively flat and sharp Moho beneath the Tarim Basin. This feature along with the large Moho offset beneath the South‐Boundary Fault suggests that the shortening and thickening of the Tien Shan crust rather than the underthrusting of the Tarim Basin are responsible for the uplift of the SCTS. In the North‐Central Tien Shan, however, the imaged Moho doublet provides direct evidence for the underthrusting of the Kazakh Shield accommodating the convergence there. Plain Language Summary: The formation of the Tien Shan mountains, situated ~1,500 km north of the plate boundary between India and Eurasia, has puzzled scientists for decades. Tectonic activities at depth could deform and thus leave footprints on subsurface interfaces such as the crust‐mantle boundary (Moho). In this study, we reveal distinct Moho structure in the southern and northern parts of the Central Tien Shan through analyses of seismic records, suggesting different mountain building processes in the two regions. The strong Tarim Basin to the south squeezes the weak South‐Central Tien Shan mass just as a wooden block squeezes plasticine. The Kazakh lower crust in the north, on the other hand, underthrusts beneath the North‐Central Tien Shan. The combined effects of the two processes lead to the rapid growth of the present‐day Tien Shan. Key Points: Detailed crustal structure beneath the Central Tien Shan is imaged, suggesting varying deformation processes perpendicular to its strikeThe shortening and thickening of the Tien Shan crust are responsible for the high elevations in the South‐Central Tien ShanThe underthrusting of the Kazakh Shield beneath the North‐Central Tien Shan accommodates the convergence there [ABSTRACT FROM AUTHOR]

Objective To investigate the effect of hemolysis on glycated albumin (GA) levels, as determined by the ketamine oxidase method. Methods GA levels and the hemolysis index were determined in nonhemolyzed serum and hemolyzed serum from corresponding patients. We developed an equation to correct the interference of hemolysis on GA, using multiple regression analysis. Results The degree of hemolysis was negatively correlated with GA levels (R2 = 0.9500). A correction equation for GA (corrected GA = 2.703 × OD of hemolysis + 1.044 × measured GA −0.906) can revert GA concentrations of hemolyzed specimens to values that were not significantly different from the GA concentration of corresponding nonhemolyzed specimens. The bias of GA concentrations before and after correction was statistically significantly different (P <.01). Conclusions Our results indicate that the level of GA measured through the ketamine oxidase method is negatively affected by hemolysis. The individualized correction of GA results provides increased accuracy in hemolyzed specimens. [ABSTRACT FROM AUTHOR]

MiR-873/CDK3 has been shown to play a critical role in ERα signaling and tamoxifen resistance. Thus, targeting this pathway may be a potential therapeutic approach for the treatment of ER positive breast cancer especially tamoxifen resistant subtype. Here we report that Norcantharidin (NCTD), currently used clinically as an ani-cancer drug in China, regulates miR-873/CDK3 axis in breast cancer cells. NCTD decreases the transcriptional activity of ERα but not ERβ through the modulation of miR-873/CDK3 axis. We also found that NCTD inhibits cell proliferation and tumor growth and miR-873/CDK3 axis mediates cell proliferation suppression of NCTD. More important, we found that NCTD sensitizes resistant cells to tamoxifen. NCTD inhibits tamoxifen induced the transcriptional activity as well ERα downstream gene expressions in tamoxifen resistant breast cancer cells. In addition, we found that NCTD restores tamoxifen induced recruitments of ERα co-repressors N-CoR and SMRT. Knockdown of miR-873 and overexpression of CDK3 diminish the effect of NCTD on tamoxifen resistance. Our data shows that NCTD regulates ERα signaling and tamoxifen resistance by targeting miR-873/CDK3 axis in breast cancer cells. This study may provide an alternative therapy strategy for tamoxifen resistant breast cancer. [ABSTRACT FROM AUTHOR]

Various geodynamic models, such as a mantle plume, lithospheric delamination and the subduction of palaeo-Pacific Plate, have been proposed to account for the Mesozoic magmatic activity and related mineralization in SE China. However, previous seismic velocity models are insufficient in resolution to constrain the accompanying crust–mantle interactions, which limits our understanding of the deep geodynamic processes responsible for the Mesozoic magmatism in SE China. In this study, we construct a new 3-D shear wave velocity model of the crust and upper mantle beneath SE China using joint inversion of receiver functions and Rayleigh-wave dispersion. We also determine the crustal thicknesses and Vp / Vs ratios by H – k stacking of receiver functions. Compared to the western Cathaysia Block, the coastal volcanic-intrusive belt (CVIB) southeast of the Zhenghe-Dapu Fault shows relatively high-velocity anomalies in the lower crust and high Vp / Vs ratios, indicating significant underplating of mafic material beneath the CVIB. Furthermore, it is noticeable that the asthenospheric low-velocity anomalies are stronger beneath the CVIB than beneath the interior of SE China, suggesting a southeastward increase in the intensity of asthenospheric upwelling and partial melting, which was likely responsible for the strong volcanism and the inferred mafic underplating beneath the CVIB. Our observations provide new evidence and more details for the northwestward subduction and subsequent rollback of the palaeo-Pacific Plate causing the Late Mesozoic magmatism in SE China. [ABSTRACT FROM AUTHOR]

• We make an exten sion of our previous wok and design a more powerful end to end n etwork for weakly supervised semantic segmentation • We propose two new loss functions for utilizing the reliable labels, including a new dense energy loss and a batch based class distance loss. The former relies on shallow features, whilst the latter focuses on distinguishing high level s emantic features for different classes. • We design a new attention module to extract comprehensive global information. By using a re weighting technique, it can suppress dominant or noisy attention values and aggregate sufficient global information. • Our approach achieves a new state of the art performance for weakly supervised semantic segmentation. Weakly supervised semantic segmentation is a challenging task that only takes image-level labels as supervision but produces pixel-level predictions for testing. To address such a challenging task, most current approaches generate pseudo pixel masks first that are then fed into a separate semantic segmentation network. However, these two-step approaches suffer from high complexity and being hard to train as a whole. In this work, we harness the image-level labels to produce reliable pixel-level annotations and design a fully end-to-end network to learn to predict segmentation maps. Concretely, we firstly leverage an image classification branch to generate class activation maps for the annotated categories, which are further pruned into tiny reliable object/background regions. Such reliable regions are then directly served as ground-truth labels for the segmentation branch, where both global information and local information sub-branches are used to generate accurate pixel-level predictions. Furthermore, a new joint loss is proposed that considers both shallow and high-level features. Despite its apparent simplicity, our end-to-end solution achieves competitive mIoU scores (val : 65.4%, test : 65.3%) on Pascal VOC compared with the two-step counterparts. By extending our one-step method to two-step, we get a new state-of-the-art performance on the Pascal VOC 2012 dataset(val : 69.3%, test : 69.2%). Code is available at: https://github.com/zbf1991/RRM. [ABSTRACT FROM AUTHOR]

Abstract: The aim of the study was to investigate the anti-asthmatic effects of apigenin and the possible mechanisms. Asthma model was established by ovalbumin-induced asthma. A total of 50 mice were randomly assigned to six experimental groups: control, model, dexamethasone (2mg/kg) and apigenin (5mg/kg, 10mg/kg). Airway resistance (Raw) was measured, histological studies were evaluated by hematoxylin and eosin (HE) staining, OVA-specific serum and BALF IgE levels and Th17 cytokines were evaluated by enzyme-linked immunosorbent assay (ELISA), Th17 cells were evaluated by flow cytometry (FCM), and protein level of RORγt was measured by western blotting. Our study demonstrated that apigenin inhibited OVA-induced increases in Raw and eosinophil count; interleukin (IL)-6, TNF-ɑ and IL-17A levels were recovered. Histological studies demonstrated that apigenin substantially inhibited OVA-induced eosinophilia in lung tissue and airway tissue. Flow cytometry studies demonstrated that apigenin substantially inhibited Th17 cells. Western blotting studies demonstrated that apigenin substantially inhibited RORγt protein level. These findings suggest that apigenin may effectively ameliorate the progression of asthma and could be used as a therapy for patients with allergic asthma. [Copyright &y& Elsevier]

Scribble annotations have recently become popular in video salient object detection. Previous methods only focus on utilizing shallow feature consistency for more integral predictions. However, there is potential for consistency between cross-frame deep features to be used to help regularize saliency predictions better. Besides, we have observed that leveraging saliency predictions as pseudo-supervision signals yields notable improvements in extracting both intra-frame and cross-frame deep features. This, in turn, leads to more precise and detailed object structural information. Thus, we propose a cross-frame feature-saliency mutual reinforcing training process to assist scribble annotations for integral video saliency predictions. Specifically, we design a cross-frame feature regularization head, which leverages intra-frame and cross-frame deep feature consistency to regularize saliency predictions as auxiliary supervision. Then, to help obtain more accurate feature consistency, we design a cross-frame saliency regularization head, where predicted saliency values are used as pseudo-supervision signals to acquire better feature consistency. In this way, our cross-frame feature and saliency regularization heads can benefit from each other to help the network learn more accurately. Extensive experiments show that our method can achieve better performances than the previous best methods. The project is available at https://github.com/muchengxue0911/CFMR. • A mutual reinforcing process is proposed to learn comprehensive object structure. • Intra-/cross-frame feature consistency is deployed to supervise saliency predictions. • Predicted saliency values are used to help build a more accurate feature consistency. [ABSTRACT FROM AUTHOR]

Several models were proposed to account for the tectonic evolution of Tibetan Plateau, among which delamination of the Tibetan lithosphere is suggested to be a leading process. However, direct evidence for the removed Tibetan lithosphere in the deep mantle remains highly lacking. Here we image the mantle transition zone (MTZ) structure beneath western Tibet and Himalaya by common conversion point stacking of 25254 receiver functions (RFs) with unprecedented clarity. One striking feature is the significant depression (15–20 km) of 660‐km discontinuity beneath Himalaya, which indicates the presence of the detached Lhasa lithospheric mantle (∼26 Ma) at the bottom of the MTZ. Our results also show ∼15 km uplift of 410‐km discontinuity (d410) and small d410 amplitude beneath western Tibet, which we ascribe to the more recently delaminated Qiangtang lithosphere (∼15 Ma). Furthermore, the small dependence of d410 amplitude on frequency indicates that the MTZ is anhydrous in the study area. Plain Language Summary: The fate of the original Tibetan lithosphere during the continental collision and its role in the tectonic evolution of the plateau remain highly debated. Here we image the mantle transition zone bounded by 410‐km (d410) and 660‐km (d660) discontinuities beneath western Tibet and Himalaya, suggesting the stepwise delamination of the Tibetan lithosphere. Our results show the significantly depressed d660 beneath the Himalaya terrane, which we interpret as reflecting the delaminated Lhasa lithosphere (∼26 Ma) stagnant near the d660. Our image also reveals prominent uplift and small amplitude of d410 beneath western Tibet, indicating the presence of delaminated Qiangtang lithosphere (∼15 Ma) in the vicinity of d410. The MTZ in the study region is anhydrous through multifrequency analysis of receiver functions, which shows small variations in d410 amplitude in all the frequency bands considered here. Key Points: Detailed structure of the mantle transition zone (MTZ) beneath western Tibet and Himalaya suggests stepwise delamination of Tibetan lithosphereThe depressed 660‐km discontinuity beneath Himalaya is due to the delaminated Lhasa lithosphere (∼26 Ma) stagnant at the bottom of the MTZThe uplifted 410‐km discontinuity beneath western Tibet is caused by the delaminated Qiangtang lithosphere (∼15 Ma) [ABSTRACT FROM AUTHOR]

Fermentation is an important link in the process of cigar production, which directly determines the quality of tobacco leaves. "Humi" fermentation is a traditional method of manually adding materials, which can highlight the aroma characteristics and improve industrial availability of tobacco leaves. However, the mechanism of "Humi" fermentation is still unclear. This study was aimed to explore the effect of the "Humi" characteristic fermentation on the quality of cigar leaves, and the potential mechanisms. Cigar variety Dexue 1 was used. Three degrees of "Humi" water were prepared and water was used as control. 16 S rDNA sequencing and metabolomics analyses were carried out to determine the changes in microbes and metabolites during fermentation. "Humi" fermentation significantly improved the evaluation quality of cigar tobacco. The microbial richness and community diversity in "Humi" fermentation group were higher than that in natural fermentation group. Total 214 genera were detected, with Staphylococcus , Corynebacterium_1 , and Oceanobacillus as the dominant genera in the "Humi" fermentation group. With the extension of fermentation time, the relative abundance of Corynebacterium_1 increased, and of Oceanobacillus and Staphylococcus first increased and then decreased. Moreover, Corynebacterium_1 were significantly negatively correlated with total nitrogen, protein and alkaloid content, and significantly positively correlated with amino acids, starch, reducing sugars and total sugars. Metabolome analysis revealed 2759 known metabolites in 36 samples and "Humi" fermentation increased the nitrogenous heterocyclic compounds, sugars, alkaloids and organic acids in tobacco leaves. These metabolites were closely related to the bacterial community. Our study provides a theoretical basis for inheriting and optimizing the "Humi" characteristic fermentation technology. [Display omitted] • "Humi" fermentation significantly improved cigar tobacco evaluation quality. • "Humi" fermentation increased the nitrogenous heterocyclic compounds. • The microbial richess and community diversity in "Humi" fermentation were higher than that in natural fermentation. [ABSTRACT FROM AUTHOR]

Deep learning-based object detection methods have achieved promising results in normal scenarios, while such methods often fail to locate objects from the disturbed images captured in inclement environments. Most existing methods utilized denoise modules to assist the detection network or exploit prior knowledge to reduce the environment interference remaining in the features, which ignores the essential role of the frequency element. From a novel perspective, we observe that the inclement environment alters the frequency content in the features, which in turn crushes the detection. To tickle this problem, we present the Magnitude Consistency Network (MCNet) to distill the irrelevant contents in the frequency domain. The MCNet is composed of the detection network and the magnitude corrector. The detection network is able to locate and classify objects. However, in inclement environments, the crucial information about the objects of interest is disrupted by the nuisance noise introduced from the inclement environments. The magnitude corrector can recover relevant information about the objects in the frequency domain by distilling the irrelevant factors and refining the affected features. Alternately optimizing the magnitude corrector and the detection network gradually makes the frequency content between the disturbed image and the clear image to be consistent. By distilling the irrelevant noise in the feature, the detection network can learn domain-invariant representations. Extensive experiments prove that the proposed method is effective and outperforms existing methods by a clear margin on four datasets. • We observe that compared to the clear images, the frequency content is altered in the inclement environment obviously, which causes the performance degradation of normal object detection approach • We propose a novel Magnitude Consistency Network (MCNet) to alleviate the distinction in the frequency components among the clear features and the disturbed features. • The experimental results demonstrate the proposed MCNet is supreme to the existing methods on the synthetic fog, synthetic rainy, natural fog and natural underwater datasets. [ABSTRACT FROM AUTHOR]

The long-tailed characteristic leads to a significant performance drop for various models on long-tailed distribution datasets. Existing works mainly focus on mitigating the data shortage in tail classes at dataset level by data re-sampling, loss re-weighting or knowledge transfer from head to tail. While in this paper, we focus on another perspective which is also related to the performance drop: the gap between total dataset class number and training batch size. To address this issue, we propose a Weight-Guided (WG) loss which utilizes the classifier weights as auxiliary tail samples. It can be easily deployed to different models. By simply adding WG loss to Mask R-CNN with ResNet-50 backbone, we improve the performance by (i) 0.5 box AP and 0.4 mask AP on COCO dataset, (ii) 0.4 box and mask AP (1.8 mask AP for rare classes) on LVIS v1.0 dataset. Codes will be released. • Point out the small class ratio covered in one batch can cause gradient shifting. • Propose a weight-guided loss to overcome the limitation of small batch size. • Our WG loss improves various frameworks consistently, especially for rare classes. [ABSTRACT FROM AUTHOR]

Video object segmentation, aiming to segment the foreground objects given the annotation of the first frame, has been attracting increasing attentions. Many state-of-the-art approaches have achieved great performance by relying on online model updating or mask-propagation techniques. However, most online models require high computational cost due to model fine-tuning during inference. Most mask-propagation based models are faster but with relatively low performance due to failure to adapt to object appearance variation. In this paper, we are aiming to design a new model to make a good balance between speed and performance. We propose a model, called NPMCA-net, which directly localizes foreground objects based on mask-propagation and non-local technique by matching pixels in reference and target frames. Since we bring in information of both first and previous frames, our network is robust to large object appearance variation, and can better adapt to occlusions. Extensive experiments show that our approach can achieve a new state-of-the-art performance with a fast speed at the same time (86.5% IoU on DAVIS-2016 and 72.2% IoU on DAVIS-2017, with speed of 0.11s per frame) under the same level comparison. Source code is available at https://github.com/siyueyu/NPMCA-net. • NPMCA-net strikes a good balance between accuracy and running speed. • Our method aggregates multi-frame information with high level interpretability. • Our methods achieves new state-of-the-art performances on reported datasets. [ABSTRACT FROM AUTHOR]

NJ001 is a monoclonal antibody that can specifically recognize the SP70 antigen on lung adenocarcinoma cells. The goal of this study was to explore its utility in targeted imaging. Subcutaneous xenograft and orthotopic lung tumor implantation BALB/c mouse models were established. Near-infrared fluorescent CF750-labeled NJ001 was injected into two tumor mouse models. Mice that received orthotopic lung tumor implantation were also injected with NJ001-conjugated nanomagnetic beads intravenously, and then underwent micro-CT scanning. Meanwhile, mice with lung tumor were intravenously injected with normal saline and bare nanomagnetic beads as a control. Fluorescence could be monitored in the mice detected by anti-SP70 fluorescence imaging, which was consistent with tumor burden. Signal intensities detected with SP70-targeted micro-CT scans were greater than those in control mice. More importantly, orthotopic tumor lesions could be found on the fourth week with SP70-targeted imaging, which was 2 weeks earlier than detection in the control. Our results suggest that SP70 is a promising target for molecular imaging, and molecularly targeted imaging with an NJ001-labeled probe could be applied for the early detection of lung adenocarcinoma. [ABSTRACT FROM AUTHOR]

• We propose a weight-guided class complementing framework to mitigate the gradient shift issue caused by unsampled classes in long-tailed scenario, including a slot-based class complementing strategy and a weight-guided feature mining loss. • We further introduce a weight refining scheme at the end of training process, which can remove the bias term in our weight-guided class complementing framework caused by long-tailed distribution. • Our plug-and-play weight-guided class complementing framework can be easily implemented to different existing approaches, achieving significant improvement on various benchmarks with new state-of-the-art performances. Real-world data are often long-tailed distributed and have plenty classes. This characteristic leads to a significant performance drop for various models. One reason behind that is the gradient shift caused by unsampled classes in each training iteration. In this paper, we propose a W eight- G uided C lass C omplementing framework to address this issue. Specifically, this framework first complements the unsampled classes in each training iteration by using a dynamic updated data slot. Then, considering the over-fitting issue caused by class complementing, we utilize the classifier weights as learned knowledge and encourage the model to discover more class specific characteristics. Finally, we design a weight refining scheme to deal with the long-tailed bias existing in classifier weights. Experimental results show that our framework can be implemented upon different existing approaches effectively, achieving consistent improvements on various benchmarks with new state-of-the-art performances. Codes will be released. [ABSTRACT FROM AUTHOR]

Despite a growing body of evidence suggests that abnormal iron metabolism plays an important role in the pathogenesis of Parkinson’s disease (PD), few studies explored its role in non-motor symptoms (NMS) of PD. The present study aimed to investigate the relationship between abnormal iron metabolism and NMS of PD. Seventy PD patients and 64 healthy controls were consecutively recruited to compare serum iron, ceruloplasmin, ferritin, and transferrin levels. We evaluated five classic NMS, including depression, anxiety, pain, sleep disorder, and autonomic dysfunction in PD patients using the Hamilton Depression Scale (HAMD), the Hamilton Anxiety Scale (HAMA), the short form of the McGill Pain Questionnaire, the Pittsburgh Sleep Quality Index and the Scale for Outcomes in Parkinson’s disease for Autonomic Symptoms, respectively. Hierarchical multiple regression analysis was used to investigate the correlations between abnormal iron metabolism and NMS. No differences in serum ceruloplasmin and ferritin levels were examined between PD patients and healthy controls, but we observed significantly decreased serum iron levels and increased serum transferrin levels in PD patients in comparison with healthy controls. After eliminating confounding factors, HAMD scores and HAMA scores were both negatively correlated with serum iron levels and positively correlated with serum transferrin levels. In summary, abnormal iron metabolism might play a crucial role in the pathogenesis of depression and anxiety in PD. Serums levels of iron and transferrin could be peripheral markers for depression and anxiety in PD. [ABSTRACT FROM AUTHOR]