Search

Single-view three-dimensional (3D) object reconstruction has always been a long-term challenging task. Objects with complex topologies are hard to accurately reconstruct, which makes existing methods suffer from blurring of shape boundaries between multiple components in the object. Moreover, most of them cannot balance learning between global geometric structure information and local detail information. In this article, we propose a multi-scale edge-guided learning network (MEGLN) to utilize the global edge information guiding the network to better capture and recover local details. The goal is to exploit the multi-scale learning strategy to learn global edge information and local details, thus achieving robust 3D object reconstruction. We first design a multi-scale Gaussian difference block (MGDB) to extract global edge geometry features for input images of different scales and adopt the attention mechanism to aggregate the extracted global edge geometry features of different scales. Second, we design a multi-scale feature interaction block (MFIB) to learn local details, which utilizes the multi-scale feature interaction to capture the features of multiple objects or components at multiple scales. The MFIB can learn and capture better as much local detail information as possible under the guidance of global edge information. Finally, we dynamically fuse the predicted probabilities of the MGDB and MFIB to obtain the final predicted result, which makes our MEGLN able to recover 3D shapes with global complex topological structures and rich local details via the multi-scale learning strategy. Extensive qualitative and quantitative experimental results on the ShapeNet dataset demonstrate that our approach achieves competitive performance compared with state-of-the-art methods. Code is available at https://github.com/Ray-tju/MEGLN .

At present, the publishing industry in China has entered the stage of knowledge-based service in digital transformation. This article introduces the four stages of the digital transformation in the Chinese publishing industry with a focus on knowledge-centered service, including the origin and meaning of the concept, the service transformation of traditional Chinese publishing industry, and the knowledge management of Internet companies and new media companies. In China, knowledge-based service of the publishing industry usually refers to a brand-new service model that integrates content creation, production, dissemination, and even application. In response to the impact of the Internet and seize the development opportunities brought by big data and artificial intelligence, the Chinese publishing industry, from traditional publishing houses to Internet-based We-Media, is actively exploring various ways and models of knowledge-based service transformation. This article illustrates the current knowledge-based services of the publishing industry in China with the case study of some pilot projects and paying for knowledge online.

In this paper, we propose a similarity-aware deep adversarial learning (SADAL) approach for facial age estimation. Instead of making full access to the limited training samples which likely leads to bias age prediction, our SADAL aims to seek batches of unobserved hard-negative samples based on existing training samples, which typically reinforces the discriminativeness of the learned feature representation for facial ages. Motivated by the fact that age labels are usually correlated in real-world scenarios, we carefully develop a similarity-aware function to well measure the distance of each face pair based on the age value gaps. Consequently, the age-difference information is exploited in the synthetic feature space for robust age estimation. During the learning process, we jointly optimize both procedures of generating hard negatives and learning discriminative age ranker via a sequence of adversarial-game iterations. Another major issue lies on that existing methods only enforce the indiscriminativeness within each class, which is probably trapped into model overfitting and thus the generation capacity is limited particularly on unseen age classes with many individuals. To circumvent this problem, we propose a variational deep adversarial learning (VDAL) paradigm, which learns to encode each face sample in two factorized parts, i.e., the intra-class variance distribution and the intra-class invariant class center. Moreover, our VDAL principally optimizes the variational confidence lower bound on the variational factorized feature representation. To better enhance the discriminativeness of the age representation, our VDAL further learns to encode the ordinal relationship among age labels in the reconstructed subspace. Experimental results on folds of widely-evaluated benchmarking datasets demonstrate that our approach achieves promising performance in contrast to most state-of-the-art age estimation methods.

Objective The aim of the present work was to investigate the expression of nitric oxide synthase 2 (iNOS/ NOS2) in colorectal and gastric cancers and evaluate its association with patient’s prognosis by integrated bioinformatics analysis. Methods The data for present study was obtained from the TCGA, GTEx, and STRING database. iNOS/NOS2 mRNA expression in normal tissue and colorectal, and gastric cancer tissuea were investigated through the GTEx and TCGA database. iNOS/NOS2 gene mutations and frequency were analyzed in the TCGA database using the cBioPortal online data analysis tool. The protein-protein interaction (PPI) network of iNOS/NOS2 was constructed by STRING database. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway of iNOS/NOS2 and relevant proteins involved in the PPI network were enriched and demonstrated by the bubble plot. Comparison of the overall survival(OS) and disease free survival(DFS) between samples expressing high and low levels of iNOS/NOS2 was analysis based on the TCGA databases through the GEPIA online data analysis tool. Results For colon adenocarcinoma (COAD) and rectal adenocarcinoma(READ) iNOS/NOS2 mRNA expression levels in tumor tissue were significant higher than those of corresponding normal colorectal tissue (ppearson=0.56, ppearson=-0.41, ppearson=0.63,ppearson=-0.43,p0.05). Conclusions iNOS/NOS2 mRNA isup-regulated in tumor tissue compared to corresponding normal tissue in colorectal and gastric cancer which implement it in the development of colorectal and gastric cancers.

Background: With the rising prevalence in recent years, gestational diabetes mellitus has become one of the leading causes of maternal and child mortality and morbidity worldwide and has raised health concern. It is seriously detrimental to both the women and fetuses. However, there are limited evidences of two types of gestational diabetes mellitus on clinical characteristics and outcomes. Therefore, this study was aimed to explore the clinical characteristics and outcomes of patients with overt diabetes mellitus (ODM) and gestational diabetes mellitus (GDM) at the late pregnancy. Methods: From January 2015 to August 2016, totally 63 gestational diabetes mellitus from the Department of Clinical Nutrition in Beijing Anzhen Hospital were enrolled in the study. Patients were classified into two groups. 31 patients with gestational overt diabetes mellitus were grouped into ODM group and 32 patients with gestational diabetes mellitus were grouped into GDM group. Clinical characteristics and outcomes were compared between ODM and GDM. We collected records of the age, gestational week, family history, past history, pregnancy complications, insulin use, blood pressure, clinical nutrition indexes, blood pressure. Glycosylated hemoglobin (HbA1c), fasting blood glucose (FBG), total protein (TP), albumin (ALB), prealbumin (PALB), hemoglobin (HGB), urea nitrogen (BUN), serum creatinine (CREA), and dynamic blood glucose monitoring were measured. And we recorded the changes of blood glucose and the test data. We statistically analyzed the data of two groups. Results: In the ODM group, HbA1c, FBG, average blood glucose, two-hour postprandial blood glucose (2hPBG) after breakfast, 2hPBG after dinner, the number of hyperglycemic events and high blood glucose time ratio are significantly higher than those of GDM and two groups compared with statistical significance (P 0.05). The number of patients treated with insulin (10/31) in ODM is significantly more than that in GDM (1/32) (P 0.05). 45% (14/31) of ODM have a family history of diabetes patients. The ratio is significantly higher than 13% (4/32) of GDM (P 0.05). There was significant difference in urinary ketone positive rate between the two groups (P 0.05), but there was no significant difference in urinary microalbumin abnormal rate between them (P > 0.05). The number of preeclampsia in ODM (8/31) is significantly higher than that of GDM (P 0.05). The level of HGB in ODM is lower than that of GDM (P 0.05). There was no difference in the pregnancy outcomes between the two groups. Conclusion: Late pregnancy women with ODM have obvious family history, higher HbA1c, higher FBG, higher glucose levels of two-hours after breakfast and dinner, higher average blood glucose, longer hypoglycemia time, higher probability of hyperglycemic events and greater opportunity to use insulin in the treatment of symptomatic patients, higher risk of preeclampsia, lower HGB level than GDM, while GDM has higher positive rate of urine ketone than ODM.

ds01: major and trace compostions of Chang'E-5 glass spherules and exterior adhensions. Extend Data Set S1: SEM micrographs of 152 Chang'E-5 glass particles analyzed in this study. No. 30 glass particle was lost during processing. Mov01-02: 3D recontruction animations of two Chang'E-5 glass spherules. Mov03: Slice animations of eight Chang'E-5 glass spherules, and arranged in No. 31, 145, 140, 4, 14, 16, 147 and 44 from top to bottom.

3D face reconstruction based on a single image is a longstanding challenging problem in computer vision. Existing end-to-end methods are difficult to reconstruct rich 3D face details. To solve this problem, we propose a two-stream convolutional neural network combined with a face super-resolution method, which can effectively restore the image’s 3D position information. Our method combines an attention fusion mechanism, which can learn the individual attention mapping of each feature subspace, and effectively learn cross-channel information while learning multi-scale and multi-frequency features. Meanwhile, our module obtains the most discriminative features in different local areas, and enhances the consistency and correlation between the attention areas. Experimental results show that our SRCNet has made significant improvements in the 3D face reconstruction and face alignment of the AFLW2000-3D and AFLW datasets.

In this paper, we propose a multi-granularity feature interaction and relation reasoning network (MFIRRN) which can recover a detail-rich 3D face and perform more accurate dense alignment in an unconstrained environment. Traditional 3DMM-based methods directly regress parameters, resulting in the lack of fine-grained details in the reconstruction 3D face. To this end, we use different branches to capture discriminative features at different granularities, especially local features at medium and fine granularities. Meanwhile, the finer-grained branch network shares its information with the adjacent coarser-grained branch network to achieve feature interaction. Our model performs cross-granular information integration and inter-granular relationship reasoning to obtain prediction results. Extensive experiments on AFLW2000-3D and AFLW datasets demonstrate the validity of our method. The code is publicly available at https://github.com/leilimaster/MFIRRN.

This paper selects the core contents of the upcoming book Market Layout and Investment Observation of Digital Content Industry in China (2017–2018), which is a continuation of the Market Layout and Investment Observation of Digital Content Industry in China (2015) with updates in contents and methodology. Keeping an eye on the latest developments of the digital content industry, we specifically studied 12 hot segments that are widely concerned and recognized in the market, including online gaming, online education, animation, online video, live streaming, We Media, news APP, online literature, online music, short video, paid knowledge and online periodicals. We made structured and systematic assortment and analysis of the developing status, competition layout, investment and financing heat and relevant policies of these segments, and summarized the ten major trends of Chinese digital content industry. We also built an investment evaluation system to objectively evaluate the investment value and the advantages and disadvantages of each segment, with investment advice and directions. As far as the general trend is concerned, the digital content industry has kept expanding in scale, competition concentrates among the industrial leaders that carry out crossover integration, investment is more rational and contents more social and vertical, IP is the hotspot, and regulation is tightened. Besides, the segments display different developing situations. For example, short video will remain under the spotlight, but the surge of paid knowledge has subsided to reveal the pain spots, online gaming will develop in a more standard way with tighter control, and the application of 5G and VR/AR technologies in vertical contents has blazed new space for tapping the existing market. Popular directions of investment include R&D of original contents, content community or socializing, IP incubation and operation, and vertical segments.

ObjectiveTo investigate the HER2/c-erbB-2, epidermal growth factor receptor (EGFR) protein expression in gastric cancer and association with patients’ clinical pathology characteristics and prognosis.MethodsHER2/c-erbB-2 and EGFR protein expression was examined by immunohistochemical assay in gastric cancer tissue and corresponding paired normal gastric tissue of 67 patients of gastric carcinoma. The HER2/c-erbB-2, EGFR protein positive expression rate in cancer tissue and normal gastric tissue were compared. The correlation between HER2/c-erbB-2, EGFR protein positive expression and patients’ clinical pathology characteristics and survival was evaluated.ResultsThe positive expression rate of HER2/c-erbB-2 in the cancer and paired normal gastric tissues were 32.8% (22/67) and 4.5% (3/67), respectively with statistical difference (p0.05).ConclusionPositive expression of HER2/c-erbB-2 and EGFR proteins in cancer tissue was significant higher than normal gastric tissue and have significant correlation with prognosis.