The tradeoff between receptive field size and efficiency is a crucial issue in low level vision. Plain convolutional networks (CNNs) generally enlarge the receptive field at the expense of computational cost. Recently, dilated filtering has been adopted to address this issue. But it suffers from gridding effect, and the resulting receptive field is only a sparse sampling of input image with checkerboard patterns. In this paper, we present a novel multi-level wavelet CNN (MWCNN) model for better tradeoff between receptive field size and computational efficiency. With the modified U-Net architecture, wavelet transform is introduced to reduce the size of feature maps in the contracting subnetwork. Furthermore, another convolutional layer is further used to decrease the channels of feature maps. In the expanding subnetwork, inverse wavelet transform is then deployed to reconstruct the high resolution feature maps. Our MWCNN can also be explained as the generalization of dilated filtering and subsampling, and can be applied to many image restoration tasks. The experimental results clearly show the effectiveness of MWCNN for image denoising, single image super-resolution, and JPEG image artifacts removal. Conv+BN+ReLU DWT IWT 160 Conv 160 160 640 256 1024 256 160 256 256 256 256 256 256 256 256 640 160 160 160 ... 256 256 256 256 256 256 Sum Connection 1024 160

BackgroundIt has been well documented that obesity is closely associated with metabolic syndrome (MetS). Although body mass index (BMI) is the most frequently used method to assess overweightness and obesity, this method has been criticized because BMI does not always reflect true body fatness, which may be better evaluated by assessment of body fat and fat-free mass. The objective of this study was to investigate the best indicator to predict the presence of MetS among fat mass index, BMI and percentage of body fat (BF %) and determine its optimal cut-off value in the screening of MetS in practice.MethodsA cross-sectional study of 1698 subjects (aged 20–79 years) who participated in the annual health check-ups was employed. Body composition was measured by bioelectrical impedance analysis (BIA). Fat mass index (FMI) was calculated. Sex-specific FMI quartiles were defined as follows: Q1: <4.39, Q2:4.39- < 5.65, Q3:5.65- < 7.03, Q4:≥7.03,in men; and Q1:<5.25, Q2:5.25- < 6.33, Q3:6.33- < 7.93,Q4:≥7.93, in women. MetS was defined by National Cholesterol Education Program/Adult Treatment Panel III criteria. The association between FMI quartiles and MetS was assessed using Binary logistic regression. Receiver operating curve(ROC) analysis was used to determine optimal cutoff points for BMI,BF% and FMI in relation to the area under the curve(AUC),sensitivity and specificity in men and women.ResultsThe adjusted odds ratios (95% CI) for the presence of MetS in the highest FMI quartile versus lowest quartile were 79.143(21.243-294.852) for men( P < 0.01) and 52.039(4.144-653.436) for women( P < 0.01) after adjusting age, BMI, BF%, TC, LDL, CRP, smoking status and exercise status, and the odds ratios were 9.166(2.157-38.952) for men( P < 0.01) and 25.574(1.945-336.228) for women( P < 0.05) when WC was also added into the adjustment. It was determined that BMI values of 27.45 and 23.85 kg/m2, BF% of 23.95% and 31.35% and FMI of 7.00 and 7.90 kg/m2 were the optimal cutoff values to predict the presence of MetS among men and women according to the ROC curve analysis. Among the indicators used to predict MetS, FMI was the index that showed the greatest area under the ROC curve in both sexes.ConclusionsHigher FMI levels appear to be independently and positively associated with the presence of MetS regardless of BMI and BF%. FMI seems to be a better screening tool in prediction of the presence of metabolic syndrome than BMI and percentage of body fat in men and women.

The Doushantuo Formation at Weng'an in Guizhou Province, South China, is best known for animal embryo-like microfossils preserved in phosphorites. However, this unit also contains a diverse assemblage of three-dimensionally phosphatized acanthomorphic acritarchs, which are useful in the biostratigraphic subdivision and correlation of the lower–middle Ediacaran System. These acritarchs can be studied using both thin sectioning and acid maceration techniques, thus have the potential to resolve taxonomic inconsistencies between acritarchs preserved in cherts and shales. This paper presents a systematic treatment of acanthomorphs (and related spheroidal microfossils) from the Doushantuo Formation at Weng'an. More than 40 distinct species are described, including the following new species:Asterocapsoides robustusn. sp.,Knollisphaeridium?bifurcatumn. sp.,Megasphaera cymbalan. sp.,Megasphaera patellan. sp.,Megasphaera puncticulosan. sp.,Mengeosphaera eccentrican. gen. n. sp.,Papillomembrana boletiformisn. sp.,Sinosphaera variabilisn. sp.,Tanarium victorn. sp.,Tianzhushania raran. sp.,Variomargosphaeridium gracilen. sp., andWeissiella brevisn. sp. The Weng'an microfossil assemblage is dominated byMegasphaeraandMengeosphaerabut shares some taxa that are characteristic of theTianzhushania spinosabiozone and theTanarium conoideum–Hocosphaeridium scaberfacium–Hocosphaeridium anozosbiozone recognized in the Yangtze Gorges area. It may represent a transitional assemblage between these two biozones. The Weng'an microfossil assemblage also shares some elements with Ediacaran acanthomorph assemblages from Australia, Siberia, and East European Platform, indicating at least partial biostratigraphic overlap with those assemblages. Among the taxa described here,T. spinosaandH. anozosemerges as easily recognizable and widely distributed acanthomorph species whose first appearance may be used to define acanthomorph biozones for regional and global biostratigraphic correlation of lower–middle Ediacaran successions.

Few layer COFs was incorporated with rGO to provide composite cathode materials with improved energy storage capability.

Microwave selective sintering is a green and efficient strategy to construct a segregated conductive network.

No abstract

In computer vision, convolutional networks (CNNs) often adopts pooling to enlarge receptive field which has the advantage of low computational complexity. However, pooling can cause information loss and thus is detrimental to further operations such as features extraction and analysis. Recently, dilated filter has been proposed to trade off between receptive field size and efficiency. But the accompanying gridding effect can cause a sparse sampling of input images with checkerboard patterns. To address this problem, in this paper, we propose a novel multi-level wavelet CNN (MWCNN) model to achieve better trade-off between receptive field size and computational efficiency. The core idea is to embed wavelet transform into CNN architecture to reduce the resolution of feature maps while at the same time, increasing receptive field. Specifically, MWCNN for image restoration is based on U-Net architecture, and inverse wavelet transform (IWT) is deployed to reconstruct the high resolution (HR) feature maps. The proposed MWCNN can also be viewed as an improvement of dilated filter and a generalization of average pooling, and can be applied to not only image restoration tasks, but also any CNNs requiring a pooling operation. The experimental results demonstrate effectiveness of the proposed MWCNN for tasks such as image denoising, single image super-resolution, JPEG image artifacts removal and object classification. The code and pre-trained models will be given at https://github.com/lpj-github-io/MWCNNv2. Index Terms-Convolutional networks, receptive field size, efficiency, multi-level wavelet.

Silicified microfossils preserved in chert nodules of the Doushantuo Formation in the Yangtze Gorges area of South China have great potential to improve the biostratigraphic subdivision and correlation of the Ediacaran System. This potential can be realized only if solid taxonomy is available. However, a systematic treatment of these microfossils (particularly acanthomorphic acritarchs) is lacking, greatly limiting their biostratigraphic potential. This paper presents the systematic paleontology of silicified microfossils from upper Doushantuo Formation (Member III) chert nodules at three sections in the Yangtze Gorges area. More than 90 species of microfossils are described, including 66 named taxa of acanthomorphs, seven named taxa of sphaeromorphs, 12 taxa of cyanobacterial filaments and coccoids, four taxa of algal thalli, and two species of tubular microfossils. Several acritarch species, includingAppendisphaera clavan. sp.,Mengeosphaera grandispinan. sp.,M. stegosauriformisn. sp.,Leiosphaeridia, and possiblySinosphaera rupina, are shown to be multicellular organisms, consistent with the proposition that some Ediacaran acritarchs may be diapause eggs of early animals. This study supports the view that theTianzhushania spinosaacanthomorph biozone is unique to the lower Doushantuo Formation in South China (and perhaps its equivalent in northern India) and that Ediacaran acanthomorph assemblages from Australia, Siberia, and East European Platform are younger than theTianzhushania spinosabiozone. It is proposed that the first occurrence ofHocosphaeridium anozos, a species with easily recognizable morphology and wide taphonomic and geographic distributions, be used to define the second Doushantuo acanthomorph biozone succeeding theTianzhushania spinosabiozone. New taxa described in this paper include three new genera (Bispinosphaeran. gen.;Yushengian. gen.; andGranitunican. gen.) and 40 new species:Appendisphaera?brevispinan. sp.,A. clavan. sp.,A.?hemisphaerican. sp.,A. longispinan. sp.,A. setosan. sp.,Bispinosphaera peregrinan. gen. n. sp.,Crinita paucispinosan. sp.,Ericiasphaera densispinan. sp.,Hocosphaeridium dilatatumn. sp.,Knollisphaeridium denticulatumn. sp.,K. longilatumn. sp.,K. obtusumn. sp.,K. parvumn. sp.,Mengeosphaera angustan. sp.,M. bellulan. sp.,M.cf.bellulan. sp.,M. constrictan. sp.,M.?cuspidatan. sp.,M.?gracilisn. sp.,M. grandispinan. sp.,M. latibasisn. sp.,M. miniman. sp.,M. spicatan. sp.,M. spinulan. sp.,M. stegosauriformisn. sp.,M. triangularisn. sp.,M. uniformisn. sp.,Sinosphaera asteriformisn. sp.,Tanarium acusn. sp.,T. elegansn. sp.,T. longitubularen. sp.,T.?minimumn. sp.,T. obesumn. sp.,T. variumn. sp.,Urasphaera fungiformisn. sp.,U. nuptan. sp.,Yushengia ramispinan. gen. n. sp.,Granitunica mcfaddeniaen. gen. n. sp.,Osculosphaera arcelliformisn. sp., andO. membraniferan. sp.