Search

Meng-Chang Yang,1,&ast; Peng Li,1,&ast; Wen-Jie Su,1 Rong Jiang,1 Jia Deng,1 Ru-Rong Wang,2 Chao-Li Huang3 1Department of Anesthesiology, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, Chengdu, People’s Republic of China; 2Department of Anesthesiology, West China Hospital, Sichuan University, Chengdu, People’s Republic of China; 3Department of Internal Medicine, Eastern Hospital, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, Chengdu, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Ru-Rong WangDepartment of Anesthesiology, West China Hospital, Sichuan University, Chengdu, Sichuan, People’s Republic of ChinaTel +86 18140049936Fax +86 28-87393632Email wangrurong@scu.edu.cnChao-Li HuangDepartment of Internal Medicine, Eastern Hospital, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, No. 585, Damian, Honghe North Road, Longquanyi District, Chengdu, People’s Republic of ChinaTel +86 13350855681Fax +86 28-87775681Email chaoli_huang@126.comPurpose: Several randomized clinical trials (RCTs) investigated the effects of the manual placental removal on hemorrhage or other hemorrhage-related complications compared with the spontaneous placental removal during cesarean section (CS), while the results remained controversial and were inconsistent. The purpose of this meta-analysis was to quantify the pooled effects of the methods of placental removal on hemorrhage during CS.Patients and Methods: A systematic literature search was conducted using PubMed, EMBASE, Web of Science, and Google Scholar. Heterogeneity was tested by I2 statistics and Q-statistic. The random-effects model or fixed-effects model were used to calculate the pooled effect for the included studies according to heterogeneity. And the term of standardized mean difference (SMD) with 95% confidence intervals (CI) was pooled and estimated the effects across all studies.Results: A total of nine RCTs were included in this meta-analysis. Compared with spontaneous group, manual placental removal increased the amount of hemorrhage (SMD = 0.53, 95% CI [0.12, 0.94]; Z = 2.54, P = 0.011) and increased the risk of endometritis (OR = 1.84, 95% CI [1.31, 2.58]; Z = 3.52, P < 0.0001). In contrast, there was no significant difference concerning the operating time (SMD = − 0.30, 95% CI [− 0.85, 0.24]; Z = 1.09, P = 0.276), the length of hospital stays (SMD = 0.11, 95% CI [− 0.08, 0.30]; Z = 1.11, P = 0.265), and blood transfusion requirement (OR = 1.36, 95% CI [0.91, 2.04]; Z = 1.52, P = 0.129), respectively.Conclusion: Comparing with spontaneous placental removal, manual placental removal appeared to be less positive effect during CS. Because of the limitations of this meta-analysis, more high-quality RCTs are needed to confirm our findings.Keywords: cesarean section, manual placental removal, spontaneous placenta removal, hemorrhage, meta-analysis

This study aimed to compare the detection performance of [68 Ga]Ga-DOTA-FAPI-04 and [18F]FDG PET/CT in the patients with various oncological and non-oncological lesions. A total of 123 patients underwent contemporaneous [68 Ga]Ga-DOTA-FAPI-04 and [18F]FDG PET/CT were included in this prospective study. The maximum standard uptake value (SUVmax) was measured to compare oncological and non-oncological lesion uptake. The sensitivity, specificity, predictive values, and accuracy of [18F]FDG and [68 Ga]Ga-DOTA-FAPI-04 PET/CT for detecting primary, metastatic, and non-oncological lesions were calculated and compared to evaluate the detection efficacy. The study subjects consisted of 123 patients (69 men and 54 women; mean age 56.11 ± 11.94 years). Among the 102 patients with either newly diagnosed (82 patients) or previously treated solid tumor (20 patients), a total of 88 solid primary malignant tumors in 84/102 patients were detected. Two patients had two primary tumors each and 1 patient had three primary tumors. Among them, 58/102 and 43/102 patients had nodal (376 lesions) and distant metastases (406 lesions), respectively. Eight patients had hematological neoplasm. No malignant oncological diseases were detected in the remaining 13 patients. A total of 145 non-oncological lesions and benign tumors in 52/123 patients were detected incidentally. [68 Ga]Ga-DOTA-FAPI-04 PET/CT demonstrated a significantly higher uptake and detection rate for the primary (SUVmax 10.98 ± 5.83 vs. 8.36 ± 6.43, p

Facial nerve palsy (FNP) is extremely rare in Kawasaki disease (KD) and the incidence is just 0.9-1.3%. Our objective was to study the clinical features and prognosis of KD accompanied with FNP and to determine whether FNP indicated an increased risk of coronary artery lesions (CALs) in KD. An observational study was performed on 9 patients diagnosed as KD accompanied with FNP in Chengdu Women’s and Children’s Central Hospital, School of Medicine, University of Electronic Science and Technology of China between September 2012 and December 2019. Patients were followed up from September 2012 till September 2020. Clinical features, laboratory and echocardiography findings, treatments, and prognosis were analyzed retrospectively. And the clinical data of patients with FNP (group A, n = 9) were compared to matched KD patients without FNP (group B, n = 27) in order to determine whether FNP indicated an increased risk of CALs in KD. The matching factors included sex, age, days of illness at the intravenous immunoglobulin use, and classic/incomplete KD forms. Clinical data of 9 patients were analyzed in this study. Patients tended to be younger than 24 months of age (88.9%) and 55.6% were under 12 months. Median onset of FNP was day 10 of illness. Eight patients (88.9%) showed unilateral FNP. The levels of laboratory inflammatory markers (White blood cells count, neutrophil percentage, C-reactive protein, erythrocyte sedimentation rate, etc.) were significantly increased in most cases. CALs were noted in 8 of 9 patients (88.9%). Coronary artery aneurysms occurred in 4 cases (44.4%) and the remaining 4 patients showed coronary artery dilation. All patients were given intravenous immunoglobulin therapy (2 g/kg), aspirin (30-50 mg/kg/day), and short-term dexamethasone. During follow-up, both FNP and CALs were completely recovered in all patients. The duration time of FNP ranged from 10 to 130 days. And the time for CALs to normal ranged from 12 to 282 days. The occurrence of CALs was significantly higher in patients with FNP than those without FNP (88.9% vs. 25.9%, P

The mechanisms of brain injury after intracerebral hemorrhage (ICH) involve mass effect-induced primary injury and secondary injury caused by a pathologic response to the hematoma. Considerable attentions have recently been paid to the mechanisms and therapeutic strategy for secondary brain injury due to no overall benefit from early surgery compared with initial conservative treatment. However, it is unclear whether there is a causal relationship between mass effect and secondary brain injury. Here, the role of mass effect on early erythrolysis after experimental ICH was investigated based on the poly(N-isopropylacrylamide) (PNIPAM) ICH model. Autologous blood and PNIPAM hydrogel were co-injected into the right basal ganglia of rats to induce different degrees of mass effect, but with a constant hematoma. The influences of different mass effect and time courses on erythrolysis and brain damages after ICH were investigated. Furthermore, the protective effect of trehalose against erythrolysis after ICH was evaluated. The results showed that mass effect caused erythrocyte morphological change at 24 hr after ICH. The released hemoglobin was quantitatively evaluated by a polynomial concerning with the mass effect, the volume of hematoma, and the time of ICH. An obvious increase in heme oxygenase-1 (HO-1) and ionized calcium binding adaptor molecule-1 (Iba-1) expression, iron deposition, cell death, and neurological deficits was observed with increasing mass effect. Moreover, trehalose alleviated brain injury by inhibiting erythrolysis after ICH. These data demonstrated that mass effect accelerated the erythrolysis and brain damages after ICH, which could be relieved through trehalose therapy.

Objectives To assess the recall of a deep learning (DL) method to automatically detect kidney stones composition from digital photographs of stones. Materials and methods A total of 63 human kidney stones of varied compositions were obtained from a stone laboratory including calcium oxalate monohydrate (COM), uric acid (UA), magnesium ammonium phosphate hexahydrate (MAPH/struvite), calcium hydrogen phosphate dihydrate (CHPD/brushite), and cystine stones. At least two images of the stones, both surface and inner core, were captured on a digital camera for all stones. A deep convolutional neural network (CNN), ResNet-101 (ResNet, Microsoft), was applied as a multi-class classification model, to each image. This model was assessed using leave-one-out cross-validation with the primary outcome being network prediction recall. Results The composition prediction recall for each composition was as follows: UA 94% (n = 17), COM 90% (n = 21), MAPH/struvite 86% (n = 7), cystine 75% (n = 4), CHPD/brushite 71% (n = 14). The overall weighted recall of the CNNs composition analysis was 85% for the entire cohort. Specificity and precision for each stone type were as follows: UA (97.83%, 94.12%), COM (97.62%, 95%), struvite (91.84%, 71.43%), cystine (98.31%, 75%), and brushite (96.43%, 75%). Conclusion Deep CNNs can be used to identify kidney stone composition from digital photographs with good recall. Future work is needed to see if DL can be used for detecting stone composition during digital endoscopy. This technology may enable integrated endoscopic and laser systems that automatically provide laser settings based on stone composition recognition with the goal to improve surgical efficiency.

Atomic force microscope (AFM)-based nanomanufacturing, as a low-cost and easy-to-setup technique, enables high-resolution maskless nanofabrication processes for arbitrary nanopatterns, but there are strong needs to advance the nanopatterning processes. Thermal-mechanical AFM-based nanomachining process lower the normal forces to few hundreds of nanonewtons, but it requires special thermal probes. Vibration-assisted nanomachining process increases the tip lifetime but forces are still in the range of few hundreds of nanonewtons. We have innovated a system with both vibration and heated sample to achieve nanomachining process with lower normal forces. To provide cantilever selection guidance for this novel nanomachining system and to understand the effect of cantilever stiffness and sample temperature on feature dimensions, we used both soft and stiff probes for the nanomachining processes in this paper. We fabricated nanostructures with controllable shape dimensions were fabricated on 200 nm depth polymethyl methacrylate (PMMA) films by using both soft and stiff AFM probes with a few tens of nanonewtons normal forces in this novel AFM-based nanomachining system. Besides, we analyzed the effects of different machining parameters on the lithography performance, which helps to understand the machining parameters for fabricating nanopatterns of different dimensions. We also found that other than the setpoint force, the probe category and sample heating temperature significantly affect the dimensions of fabricated nanopatterns. And there is a statistically significant interaction effect between the spring constants of probes and the sample temperatures.

This article summarizes technical advances contained in the fifth major release of the Q-Chem quantum chemistry program package, covering developments since 2015. A comprehensive library of exchange–correlation functionals, along with a suite of correlated many-body methods, continues to be a hallmark of the Q-Chem software. The many-body methods include novel variants of both coupled-cluster and configuration-interaction approaches along with methods based on the algebraic diagrammatic construction and variational reduced density-matrix methods. Methods highlighted in Q-Chem 5 include a suite of tools for modeling core-level spectroscopy, methods for describing metastable resonances, methods for computing vibronic spectra, the nuclear–electronic orbital method, and several different energy decomposition analysis techniques. High-performance capabilities including multithreaded parallelism and support for calculations on graphics processing units are described. Q-Chem boasts a community of well over 100 active academic developers, and the continuing evolution of the software is supported by an “open teamware” model and an increasingly modular design. This article summarizes technical advances contained in the fifth major release of the Q-Chem quantum chemistry program package, covering developments since 2015. A comprehensive library of exchange-correlation functionals, along with a suite of correlated many-body methods, continues to be a hallmark of the Q-Chem software. The many-body methods include novel variants of both coupled-cluster and configuration-interaction approaches along with methods based on the algebraic diagrammatic construction and variational reduced density-matrix methods. Methods highlighted in Q-Chem 5 include a suite of tools for modeling core-level spectroscopy, methods for describing metastable resonances, methods for computing vibronic spectra, the nuclear-electronic orbital method, and several different energy decomposition analysis techniques. High-performance capabilities including multithreaded parallelism and support for calculations on graphics processing units are described. Q-Chem boasts a community of well over 100 active academic developers, and the continuing evolution of the software is supported by an "open teamware" model and an increasingly modular design.

Supplemental Digital Content is available in the text, Objective: Endometriosis is a common gynecologic disease that frequently leading to chronic pelvic pain, severe dysmenorrhea, and subfertility. As first-line hormonal treatment can interfere with ovulation and may cause recurrent pelvic pain, exploration of new non-hormonal therapeutic approaches becomes increasingly necessary. This review aimed to evaluate the pre-clinical and clinical efficacy and safety of non-hormonal treatment for endometriosis Data sources: Databases including PubMed, Embase, Cochrane Library, SINOMED, ClinicalTrials.gov, and Google Scholar were searched up to October 2019, using search terms “endometriosis” and “non-hormonal therapy.” Study selection: Twenty-four articles were reviewed for analysis, including nine animal studies and 15 human trials; all were published in English. Results: Twenty-four articles were identified, including 15 human trials with 861 patients and nine animal studies. Some agents have been evaluated clinically with significant efficacy in endometriosis-related pelvic pain and subfertility, such as rofecoxib, etanercept, pentoxifylline, N-palmitoylethanolamine, resveratrol, everolimus, cabergoline (Cb2), and simvastatin. Other drugs with similar pharmacological properties, like parecoxib, celecoxib, endostatin, rapamycin, quinagolide, and atorvastatin, have only been tested in animal studies. Conclusions: Clinical data about most of the non-hormonal agents are not sufficient to support them as options for replacement therapy for endometriosis. In spite of this, a few drugs like pentoxifylline showed strong potential for real clinical application.

Surgery is an important therapeutic strategy for intracerebral hemorrhage (ICH) in the clinic and is theoretically beneficial for the outcome of ICH by decreasing hematoma, reducing nervous tissue damage, and removing harmful chemicals. However, the outcome of ICH surgery is always unsatisfactory due to postoperative rebleeding. We hypothesized that the injection of hemostatic agents in situ after aspiration surgery could immediately activate hemostasis once rebleeding occurs. Therefore, keratin hydrogels (K-gels) were easily prepared as a hemostatic material via a rehydration method and had a porous structure. Collagenase was injected into the basal lamina to mimic ICH rebleeding, and the K-gels were then injected into the same injured site after 2 h for hemostatic therapy. The hematoma volume was significantly reduced by K-gel treatment, indicating that in situ infusion of the K-gels inhibited hematoma enlargement when rebleeding occurred. Moreover, brain damage, including cell apoptosis, neuroinflammatory reactions, and neurological deficits, was also relieved after K-gel treatment. These results suggested that in situ injection of the K-gels into the hematoma area after ICH surgery improves the therapeutic outcome by stopping postoperative rebleeding. K-gels have great potential for clinical hemostatic application because of their excellent hemostatic properties and biocompatibility.

The recognition of human emotions from facial expression images is one of the most important topics in the machine vision and image processing fields. However, recognition becomes difficult when dealing with non-frontal faces. To alleviate the influence of poses, we propose an encoder-decoder generative adversarial network that can learn pose-invariant and expression-discriminative representations. Specifically, we assume that a facial image can be divided into an expressive component, an identity component, a head pose component and a remaining component. The encoder encodes each component into a feature representation space and the decoder recovers the original image from these encoded features. A classification loss on the components and an ℓ1 pixel-wise loss are applied to guarantee the rebuilt image quality and produce more constrained visual representations. Quantitative and qualitative evaluations on two multi-pose datasets demonstrate that the proposed algorithm performs favorably compared to state-of-the-art methods.

Bio-mimic wings are crucial constituents of Flapping-Wing Micro Air Vehicles, and the quality of such products plays a vital role in missions involving area exploration, aerial imaging and monitoring. Fused filament fabrication, as a versatile manufacturing approach, can be adopted to print single-layer artificial wings in an additive way. In such processes, quality of the printed wings is determined by the percentage of covered area, contingent on the vertical distance between the printing nozzle and the substrate. Even when properly calibrated, this distance is subject to variations due to the exceeding sensitivity to process drift, including vibration of the platform. In situ image analysis provides rapid guidance for process drift detection and assessment of the printed wing structures. While dynamic network analysis has been applied to quality assurance of steaming images, the computational overhead has made it unrealistic for monitoring high-definition images, particularly in the detection of images with incipient or tiny defects. In this study, we investigate image decomposition techniques, including wavelet transformation and bi-dimensional empirical mode decomposition (BEMD), to dissolve the original images into a series of basis images, which significantly reduces the number of communities in the network representation of images and diminishes the computational cost. Lastly, network generalized likelihood ratio control chart is developed to characterize anomalous variations in structures of the images. Our investigation shows that wavelet and EMD elevate the detection accuracy by 36.7% and 63.3%, while reducing the processing time by 43.5% and 73.9%, respectively, compared to the dynamic network approach without decomposition.

Microlens arrays (MLAs), as one of the key features in the optoelectronics field, have attracted a lot of attention recently. Unfortunately, existing fabrication methods of MLAs have many disadvantages, such as complex fabrication procedures and difficult morphology control. This paper presents a low-cost and tunable fabrication approach for concave MLAs on ultraviolet (UV) polymer surfaces. We used condensed water droplets, formed by initiative cooling, as the template for the formation of concave MLAs in this approach. A sacrificial layer of polymer material was introduced to fabricate concave MLAs with different morphology parameters. The three most important parameters of MLAs, i.e. diameter, cross-sectional profile, and packing distance, can be tuned by the proposed novel fabrication approach. By controlling the condensation time of water droplets, we can control the diameter of the concave MLAs. Moreover, the cross-sectional profile can be controlled by replacing the sacrificial layer material to change the interaction among the water droplets, sacrificial polymers, and UV polymers. In addition, the packing distance of the MLAs can also be adjusted by introducing the lateral flow of the sacrificial polymer, which was driven by the additional sacrificial polymer dispensed on the substrate. Furthermore, the UV polymer film with MLAs on its surface was applied for the packaging of high power green LEDs. Consequently, the optical output power of green LED modules is enhanced by 11.4% when using the MLAs at the driving current of 500 mA.

Objective: This study aimed to compare the diagnostic performance of 68Ga-FAPI-04 and 18F-FDG PET/CT in the patients with various oncological and non-oncological lesions. Patients and Methods: A total of 123 patients underwent contemporaneous 68Ga-FAPI-04 and 18F-FDG PET/CT were included in this prospective study. The maximum standard uptake value (SUVmax) was measured to compare oncological and non-oncological lesion uptake. The sensitivity, specificity, predictive values and accuracy of 18F-FDG and 68Ga-FAPI-04 PET/CT for detecting primary, metastatic, and non-oncological lesions were calculated and compared to evaluate the diagnostic efficacy. Results: The study subjects consisted of 123 patients (69 men and 54 women; mean age: 56.11±11.94). A total of 84 patients with 88 solid primary malignant tumors, 58 patients with 376 nodal metastases, 43 patients with 406 distant metastases, 8 patients with hematological neoplasms and 52 patients with 145 non-oncological lesions and benign tumors were detected. 68Ga-FAPI-04 PET/CT demonstrated a significantly higher uptake and detection rate for the primary (SUVmax: 10.98±5.83 vs. 8.36±6.43, p＜0.001; X2=0.538, p=0.021), nodal (SUVmax: 10.50±5.98 vs. 8.20±6.29, p=0.011; X2=2.067, p＜0.001) and distant metastatic lesions (SUVmax: 6.74±4.83 vs. 9.64±6.45; p＜0.001; X2=4.897, p＜0.001) of solid tumor than did 18F-FDG PET/CT. 68Ga-FAPI-04 PET/CT demonstrated a lower activity (SUVmax: 6.84±4.67 vs. 13.09±7.29, p＜0.001) and detection rate (X2=5.166, p＜0.001) for multiple myeloma and lymphoma compared to 18F-FDG PET/CT. 68Ga-FAPI-04 and 18F-FDG PET/CT PET/CT demonstrated a comparative diagnostic efficacy (SUVmax: 6.40±3.95 vs. 5.74±15.78, p = 0.729; X2 = 9.460, p = 0.007) for non-oncological lesion and benign tumor detection. Conclusions: Except for myeloma and lymphoma, 68Ga-FAPI-04 PET/CT showed a superior diagnostic efficacy for detecting various primary and metastatic lesions than 18F-FDG PET/CT. A comparative diagnostic utility for non-oncological lesion detection was obtained with both tracers. 68Ga-FAPI-04 could be used as a broad-spectrum tumor and inflammatory imaging agent in the clinical especially for various solid tumors and non-oncological lesions.

We address the problem of performing backpropagation for computation graphs involving 3D transformation groups SO(3), SE(3), and Sim(3). 3D transformation groups are widely used in 3D vision and robotics, but they do not form vector spaces and instead lie on smooth manifolds. The standard backpropagation approach, which embeds 3D transformations in Euclidean spaces, suffers from numerical difficulties. We introduce a new library, which exploits the group structure of 3D transformations and performs backpropagation in the tangent spaces of manifolds. We show that our approach is numerically more stable, easier to implement, and beneficial to a diverse set of tasks. Our plug-and-play PyTorch library is available at https://github.com/princeton-vl/lietorch., Comment: fixed typos

We presented a case of 67-year-old man with epigastric discomfort and anorexia for more than 2 months. No abnormal 18F-FDG uptake was found throughout the gastrointestinal tract on 18F-FDG PET/CT. The patient was enrolled in the 68Ga-FAPI PET/CT clinical trial, and a 68Ga-FAPI PET/CT was performed to help detect primary lesion. This examination demonstrated a nodular thickening of the gastric wall with an increased 68Ga-FAPI uptake in the greater curvature of the stomach. Finally, the pathological result confirmed the diagnosis of gastric stromal tumor. This case highlights that 68Ga-FAPI PET/CT might outperform 18F-FDG PET/CT in helping identify gastrointestinal stromal tumors.

We address the problem of scene flow: given a pair of stereo or RGB-D video frames, estimate pixelwise 3D motion. We introduce RAFT-3D, a new deep architecture for scene flow. RAFT-3D is based on the RAFT model developed for optical flow but iteratively updates a dense field of pixelwise SE3 motion instead of 2D motion. A key innovation of RAFT-3D is rigid-motion embeddings, which represent a soft grouping of pixels into rigid objects. Integral to rigid-motion embeddings is Dense-SE3, a differentiable layer that enforces geometric consistency of the embeddings. Experiments show that RAFT-3D achieves state-of-the-art performance. On FlyingThings3D, under the two-view evaluation, we improved the best published accuracy (d < 0.05) from 34.3% to 83.7%. On KITTI, we achieve an error of 5.77, outperforming the best published method (6.31), despite using no object instance supervision. Code is available at https://github.com/princeton-vl/RAFT-3D.

Objective To explore the characteristics of coagulase‐negative Staphylococci other than Staphylococci epidermidis (Nse‐CoNS) meningitis and to apply cerebrospinal fluid (CSF) times to positivity culture (TTPC) for the precise differentiation of meningitis from contamination. Methods We conducted a case‐control study to accomplish the following: First, we retrospectively reviewed records of post‐neurosurgical patients’ CSF that yielded Nse‐CoNS from January to October 2019 at the Beijing Tiantan Hospital; 17 clinical and 12 laboratory characteristics were reviewed. Second, we investigated the TTPC of the Nse‐CoNS, the cutoffs, and corresponding parameters to differentiate Nse‐CoNS meningitis from contamination. Results In this study, a total of 146 patients with Nse‐CoNS CSF culture positive were enrolled. The average TTPC in the Nse‐CoNS meningitis group was significantly shorter than in the contamination group (20.2 ± 5.0 hours and 30.2 ± 12.6 hours, respectively, P, Cerebrospinal fluids times to positivity culture(TTPC) can differentiate coagulase‐negative Staphylococci other thanStaphylococci epidermidis (Nse‐CoNS) meningitis from contamination. A TTPC shorter than 20.0 h was associated with Nse‐CoNS meningitis, while a TTPC longer than 20.0 h was associated with Nse‐CoNS contamination.

Synthetic images rendered by graphics engines are a promising source for training deep networks. However, it is challenging to ensure that they can help train a network to perform well on real images, because a graphics-based generation pipeline requires numerous design decisions such as the selection of 3D shapes and the placement of the camera. In this work, we propose a new method that optimizes the generation of 3D training data based on what we call "hybrid gradient". We parametrize the design decisions as a real vector, and combine the approximate gradient and the analytical gradient to obtain the hybrid gradient of the network performance with respect to this vector. We evaluate our approach on the task of estimating surface normal, depth or intrinsic decomposition from a single image. Experiments on standard benchmarks show that our approach can outperform the prior state of the art on optimizing the generation of 3D training data, particularly in terms of computational efficiency.

Single-view 3D is the task of recovering 3D properties such as depth and surface normals from a single image. We hypothesize that a major obstacle to single-image 3D is data. We address this issue by presenting Open Annotations of Single Image Surfaces (OASIS), a dataset for single-image 3D in the wild consisting of annotations of detailed 3D geometry for 140,000 images. We train and evaluate leading models on a variety of single-image 3D tasks. We expect OASIS to be a useful resource for 3D vision research. Project site: https://pvl.cs.princeton.edu/OASIS., Accepted to CVPR 2020

Background Lidocaine is a commonly used local anesthetic, and low-dose lidocaine has neuroprotective effects on cerebral ischemia/reperfusion (CI/R) injury; the mechanism for this, however, is still unclear. The aim of this study was to investigate the role and the possible mechanisms of lidocaine on CI/R injury in rats. Methods We constructed a rat (male Sprague-Dawley rats, 6-8 weeks old) model of CI/R injury induced by middle cerebral artery occlusion (MCAO). Histopathology, neuronal apoptosis, oxidative stress, and inflammatory response were evaluated using hematoxylin and eosin (HE) staining, Nissl staining, enzyme-linked immunosorbent assay (ELISA) and western blotting, respectively. In addition, brain water content, infarct volume, neurological deficit score each evaluated. Results The findings showed that lidocaine improved spatial learning and memory impairment, protected I/R-induced brain injury and attenuated neuronal death and apoptosis. Furthermore, lidocaine also regulated the levels of malondialdehyde (MDA), glutathione (GSH), superoxide dismutase (SOD), IL-6, IL-10, iNOS, and IL-4.Notably, lidocaine markedly inhibited the expression of p65 and p38. Conclusions The results indicate that lidocaine protects against cerebral injury induced by I/R in rats via the nuclear factor kappa-B (NF-κB) p65 and p38 mitogen-activated protein kinase (MAPK) signaling pathway, it provided a candidate for the treatment of CI/R-induced injury.

Objective: The aim of this study was to compare the treatment effects of high-intensity focused ultrasound (HIFU) and laparoscopic excision (LE) in patients with adenomyosis and infertility.Materials and methods: A total of 93 patients with adenomyosis and infertility who were treated with HIFU (50 patients) or LE (43 patients) from January 2012 to January 2017 at the Third Xiangya Hospital of Central South University were retrospectively analyzed. Clinical characteristics including dysmenorrhea severity pain score, menorrhagia severity scores, reproductive outcomes, complications during pregnancy and delivery, adverse effects, surgical complications, and other clinical variables were compared between the HIFU and LE groups.Results: Of the total 93 patients with adenomyosis and infertility, 50 were treated with HIFU and 43 underwent LE. Both HIFU and LE treatments achieved significant relief of dysmenorrhea and menorrhagia. The total hospital stay was shorter in patients treated with HIFU than in those who underwent LE surgery. Neither HIFU nor LE treatment led to severe complications after treatment. Most importantly, patients treated with HIFU showed significantly higher pregnancy rates and natural conception rates than those who underwent LE surgery. Notably, in the HIFU treatment group, those with diffuse adenomyotic lesions had significantly lower postoperative pregnancy rates than those with focal adenomyosis.Conclusion: HIFU showed a safe and effective profile as a therapeutic management option for patients with adenomyosis. In comparison with LE, HIFU treatment achieved better postoperative reproductive outcomes. HIFU treatment should be encouraged and implemented in clinical practice.

A lower mortality rate is observed in obese patients with acute lung injury (ALI), which is referred to as the obesity paradox, in several studies and recent meta-analyses. Hyperinsulinemia is characterized as the primary effect of obesity, and exogenous insulin attenuates LPS-induced pulmonary edema. The detailed mechanism responsible for the effect of hyperinsulinemia on pulmonary edema and alveolar filling needs to be elucidated. SD rats were fed with a high-fat diet (HFD) for a total of 14 weeks. SD rats were anesthetized and intraperitoneally injected with 10 mg/kg lipopolysaccharide (LPS), while control rats received only saline vehicle. Insulin receptor antagonist S961 (20 nmol/kg) was given by the tail vein and serum, and glucocorticoid-induced protein kinase-1 (SGK-1) inhibitor EMD638683 (20 mg/kg) was administrated intragastrically prior to LPS exposure. The lungs were isolated for the measurement of alveolar fluid clearance. The protein expression of epithelial sodium channel (ENaC) was detected by Western blot. Insulin level in serum was significantly higher in HFD rats compared with normal diet rats in the presence or absence of LPS pretreatment. Hyperinsulinemia induced by high fat feeding increased alveolar fluid clearance and the abundance of α-ENaC, β-ENaC, and γ-ENaC in both normal rats and ALI rats. Moreover, these effects were reversed in response to S961. EMD638683 prevented the simulation of alveolar fluid clearance and protein expression of ENaC in HFD rats with ALI. These findings suggest that hyperinsulinemia induced by obesity results in the stimulation of alveolar fluid clearance via the upregulation of the abundance of ENaC in clinical acute lung injury, whereas theses effects are prevented by an SGK-1 inhibitor.

OBJECTIVE The aim of the study was to examine whether an electrical impedance tomography (EIT) electrode belt changed the lung function in healthy volunteers and patients with respiratory muscle weakness (RMW) and chronic obstructive pulmonary disease (COPD). APPROACH In total, thirty subjects were included (10 healthy volunteers, 10 subjects with RMW, maximum inspiratory pressure < 40 cmH2O, and 10 COPD, grade I-IV). Spirometry measurements were conducted in a sitting position once a day at similar times on two consecutive days. Slow expiratory vital capacity (VC), forced vital capacity (FVC) and maximum voluntary ventilation (MVV) manoeuvres were performed. On day 1, spirometry was performed without the EIT electrode belt, and on day 2, the belt was attached to the thorax. MAIN RESULTS Lung function was not influenced by the electrode belt in healthy subjects. The test-retest reliability in the healthy group was 0.89, 0.89 and 0.85 for VC, FVC and MVV, respectively. On the other hand, all investigated parameters were significantly decreased in the RMW group (VC, 51.3 ± 18.0 versus 46.5 ± 18.0% predicted, without versus with EIT belt, p< 0.01; FVC, 51.7 ± 19.0 versus 45.8 ± 18.1% predicted, p< 0.01; MVV, 41.0 ± 20.0 versus 38.8 ± 19.6% predicted, p< 0.01). VC and MVV also decreased significantly in the COPD group (VC, 77.4 ± 20.5 versus 74.6 ± 18.8% predicted, p< 0.05; MVV, 57.4 ± 15.7 versus 54.4 ± 12.5% predicted, p< 0.05). SIGNIFICANCE An EIT electrode belt could reduce lung volumes in subjects with pre-existing lung diseases. Comparing lung function acquired with an electrode belt to corresponding values obtained without the belt should be avoided.

Computer vision technology is being used by many but remains representative of only a few. People have reported misbehavior of computer vision models, including offensive prediction results and lower performance for underrepresented groups. Current computer vision models are typically developed using datasets consisting of manually annotated images or videos; the data and label distributions in these datasets are critical to the models' behavior. In this paper, we examine ImageNet, a large-scale ontology of images that has spurred the development of many modern computer vision methods. We consider three key factors within the "person" subtree of ImageNet that may lead to problematic behavior in downstream computer vision technology: (1) the stagnant concept vocabulary of WordNet, (2) the attempt at exhaustive illustration of all categories with images, and (3) the inequality of representation in the images within concepts. We seek to illuminate the root causes of these concerns and take the first steps to mitigate them constructively., Accepted to FAT* 2020

Recent advances have spurred incredible progress in self-supervised pretraining for vision. We investigate what factors may play a role in the utility of these pretraining methods for practitioners. To do this, we evaluate various self-supervised algorithms across a comprehensive array of synthetic datasets and downstream tasks. We prepare a suite of synthetic data that enables an endless supply of annotated images as well as full control over dataset difficulty. Our experiments offer insights into how the utility of self-supervision changes as the number of available labels grows as well as how the utility changes as a function of the downstream task and the properties of the training data. We also find that linear evaluation does not correlate with finetuning performance. Code and data is available at \href{https://www.github.com/princeton-vl/selfstudy}{github.com/princeton-vl/selfstudy}., Comment: To appear in CVPR 2020

This paper develops and characterizes a three-dimensional (3-D) nanofabrication process using ultrasonic vibration assisted nanomachining based on an atomic force microscope (AFM). The superiorities of height control over force control in the process are explained and are demonstrated by the fabrication results. Three factors impacting actual feature depths are investigated, including the ultrasonic z -vibrational amplitude, the assigned base feature depth, and the machining speed. 3-D nanostructures with continuous height variations were successfully fabricated on polymethyl methacrylate (PMMA) films with the feature height manipulated through controlling the absolute height of the cantilever tip in AFM. By selecting machining parameters based on characterizations, feature dimensions can be controlled as desired values within small variances. The capability of transferring 3-D nanostructures from PMMA films to silicon substrates is further explored in this paper. After selecting recipes of the reactive ion etching process, 3-D nanostructures are successfully transferred to silicon substrates with controllable selectivity. The reported ultrasonic vibration assisted nanomachining process in height control provides a robust approach of fabricating 3-D nanostructures.

China is the largest rice producing and consuming country in the world, accounting for more than 25% of global production and consumption. Rice cultivation is also one of the main sources of anthropogenic methane (CH4) and nitrous oxide (N2O) emissions. The challenge of maintaining food security while reducing greenhouse gas emissions is an important tradeoff issue for both scientists and policy makers. A systematical evaluation of tradeoffs requires attention across spatial scales and over time in order to characterize the complex interactions across agricultural systems components. We couple three well-known models that capture different key agricultural processes in order to improve the tradeoff analysis. These models are the DNDC biogeochemical model of soil denitrification-decomposition processes, the DSSAT crop growth and development model for decision support and agro-technology analysis, and the regional AEZ crop productivity assessment tool based on agro-ecological analysis. The calibration of eco-physiological parameters and model evaluation used the phenology and management records of 1981–2010 at nine agro-meteorological stations spanning the major rice producing regions of China. The eco-physiological parameters were calibrated with the GLUE optimization algorithms of DSSAT and then converted to the counterparts in DNDC. The upscaling of DNDC was carried out within each cropping zone as classified by AEZ. The emissions of CH4 and N2O associated with rice production under different management scenarios were simulated with the DNDC at each site and also each 10 × 10 km grid-cell across each cropping zone. Our results indicate that it is feasible to maintain rice yields while reducing CH4 and N2O emissions through careful management changes. Our simulations indicated that a reduction of fertilizer applications by 5–35% and the introduction of midseason drainage across the nine study sites resulted in reduced CH4 emission by 17–40% and N2O emission by 12–60%, without negative consequences on rice yield.

Surface enhanced Raman spectroscopy (SERS) is a surface-sensitive detection technique that dramatically increases the scattering signals of the analytes compared to traditional Raman spectroscopy. Rapid and low-cost fabrication of SERS substrates with easily tunable features remains a challenge, although many SERS substrates with high enhancement factors (EF) were investigated. Here, we report a novel and rapid approach of fabricating SERS substrates using ultrasonic vibration assisted nanomachining. Grids patterns with easily tuned dimensions were fabricated on PMMA surfaces. SERS substrates were fabricated after coating an 80 nm gold layer on the patterned silicon after reactive ion etching (RIE) using the patterned PMMA as the mask. Probing molecules of R6G with the area density of 2.8 × 10-12 M/mm2 can be detected, achieving an EF of 3.11 × 103. Results show a pattern with higher grids density tends to achieve a higher EF.

Surface electromyography signal plays an important role in hand function recovery training. In this paper, an IoT-enabled stroke rehabilitation system was introduced which was based on a smart wearable armband (SWA), machine learning (ML) algorithms, and a 3-D printed dexterous robot hand. User comfort is one of the key issues which should be addressed for wearable devices. The SWA was developed by integrating a low-power and tiny-sized IoT sensing device with textile electrodes, which can measure, pre-process, and wirelessly transmit bio-potential signals. By evenly distributing surface electrodes over user’s forearm, drawbacks of classification accuracy poor performance can be mitigated. A new method was put forward to find the optimal feature set. ML algorithms were leveraged to analyze and discriminate features of different hand movements, and their performances were appraised by classification complexity estimating algorithms and principal components analysis. According to the verification results, all nine gestures can be successfully identified with an average accuracy up to 96.20%. In addition, a 3-D printed five-finger robot hand was implemented for hand rehabilitation training purpose. Correspondingly, user’s hand movement intentions were extracted and converted into a series of commands which were used to drive motors assembled inside the dexterous robot hand. As a result, the dexterous robot hand can mimic the user’s gesture in a real-time manner, which shows the proposed system can be used as a training tool to facilitate rehabilitation process for the patients after stroke., An IoT-enabled stroke rehabilitation system was implemented which was based on a smart wearable armband, machine learning algorithms, and a 3D printed dexterous robot hand. The smart wearable armband was developed by integrating a low-power and tiny-sized IoT sensing device with textile electrodes, which can measure, pre-process, and wirelessly transmit bio-potential signals. Machine learning (ML) algorithms were leveraged to analyze and discriminate features of different hand movements. The patient's hand movement intentions were further extracted to control the dexterous robot hand and the robot hand can mimic the patient's gesture in a real-time manner which enables active response to patient's volition.

Hepatic alveolar echinococcosis (HAE) is a severe and common parasitic disease in Tibetan Plateau of China. The infected patients have to move to plain areas to receive treatments due to the poor medical conditions in plateau areas. Our aim was to investigate the application of Enhanced Recovery after Surgery (ERAS) program in perioperative management for HAE patients from Tibet Plateau and the notes for patients with landform changes.A total of 89 HAE patients from Tibet Plateau (altitude: average of 4500 m) prior received adaptive treatments at the cooperative hospital (altitude: 1500-2000 m) and accepted surgery at plain regions (altitude: 200-400 m). The patients in ERAS group received ERAS program care and patients in conventional management group received conventional care during perioperative period.Patients in ERAS group displayed significant shorter hospital stay and shorter time for recovery of gurgling compared with conventional management group (ERAS group versus conventional management group: 10.48 ± 3.525 d versus 20.29 ± 8.632 d; 1.56 ± 1.236 d versus 2.8 ± 1.19 d; all P 0.01). The number of patients with complications of bloating, nausea/vomiting, pulmonary infection, urinary tract infection, upper gastrointestinal hemorrhage, and pulmonary edema was remarkably reduced (number, ERAS group versus conventional management group: 14 versus 24; 5 versus 16; 7 versus 24; 4 versus 13; 0 versus 10; all P 0.05), and the visual analog scale scores in postoperative days 1 and 2 were obviously decreased in patients of ERAS group (score, ERAS group versus conventional management group: 2.5 ± 1.288 versus 3.83 ± 1.87; 2.25 ± 0.838 versus 3.51 ± 1.468; all P 0.01).Patients from Tibet Plateau need to receive adaptive treatments for landform changes before receiving surgeries at plain regions. ERAS program is effective and safe for Tibetan HAE patients during perioperative period.

During the past century Chinese agriculture has been struggling to produce more food to support the ever growing population, while dealing with the increased degradation of air, water and soil quality in the agricultural regions. Lessons learnt from the long-term efforts indicated that scientifically sound methodology could play a crucial role in understanding the complex interactions among climate, soil, water and farming management practices, all of which collectively control the agroecosystem services including crop yield, greenhouse gas emissions, nutrient loading and other environmental issues. Therefore, the development of a more process-based approach is needed. The process-based model, DNDC (Denitrification-Decomposition) has been widely used internationally to simulate detailed carbon and nitrogen biogeochemical cycles occurring in agricultural systems. However, this model is not fully suitable for China as it lacks a number of features which are crucial for representing Chinese agro-ecosystems, including paddy rice cultivation, complex and multiple cropping systems, and intensive management practices, etc. Recently a new version of DNDC, China-DNDC, was developed. The new model improved its capability of predicting the fluxes of all three terrestrial greenhouse gases: nitrous oxide (N2O), carbon dioxide (CO2), and methane (CH4), as well as other important indicators such as crop phenology, soil C sequestration, and nutrient leaching from different cropping systems across all the major agricultural regions in China. This paper reported how China-DNDC was developed, tested and applied for sustaining Chinese agriculture. And then, it identified the weaknesses and potential improvements for the model.

Catenary is an important part of electrified railway, and its geometric parameters are important parameters reflecting the safe and stable operation of locomotives. With the improvement of its speed, there are higher requirements for high-accuracy and real-time detection of geometric parameters of catenary. The existing systems have problems of long sampling interval, low real-time performance, and light-sensitive. Aiming at the actual requirement of dynamic measurement of catenary geometric parameters, a non-contact catenary geometric parameter detection system based on machine vision was developed. Firstly, a measurement model based on high-power line lasers and high-resolution area cameras was established to meet the application requirements. The measurement principle of the system was analyzed and the detailed formulas were deduced. Secondly, image difference, laser spot roundness analysis and other image processing algorithms were used to quickly and accurately detect the characteristics of laser points on the contact line with complex background. Based on the measurement model and algorithms mentioned above, the hardware and software platform of the system were built, and fast image acquisition and processing was realized by using multi-thread programming technology on high-performance industrial computer, which solved the problems of long sampling interval and low real-time performance during the measurement. Real-time image storage and display and preservation of detection results were realized in the software. Finally, a preliminary experiment was performed on the prototype, and the accuracy of the measurement results was analyzed. Experiment results showed that the system works stably and has high accuracy, which meets the practical application requirements.

The contents of steroids and endocannabinoids along with the ratios between them would be candidate biomarkers for sensitively and comprehensively assessing the role of testosterone in regulating the activities of the hypothalamus-pituitary-gonadal (HPG) axis, the hypothalamus-pituitaryadrenal (HPA) axis and endogenous cannabinoid system (ECS). However, previous studies mostly used the contents rather than the ratios as biomarkers. This study aimed to systematically screen and identify sensitive biomarkers from 21 candidates including both the contents of nine steroids and one endocannabinoid and their ratios in saliva. Three screening criteria were whether there were intergroup differences, time-dependent changes and considerable relative stability during a 4-h period after exogenous testosterone administration. This study used LC-APCI+-MS/MS to determine the salivary levels of the candidate biomarkers on 62 male healthy undergraduates who were divided into testosterone administration and placebo control groups. The results revealed that salivary testosterone, androstenedione, DHEA and the ratios of testosterone to estradiol and AEA, and of cortisol to testosterone and DHEA were sensitive biomarkers for assessing the effects of testosterone administration on the three neuroendocrine systems because they all showed significant intergroup differences and time-dependent changes and good relative stability. Salivary cortisol, cortisone and the ratios of testosterone to androstenedione and DHEA and of androstenedione to estrone, and of cortisol to cortisone, androstenedione and AEA might be suitable biomarkers because they met only two of the three criteria, but needed to be validated in the future. The rest biomarkers were unsuitable because they mostly showed no significant intergroup differences, blunt time-dependent changes and poor relative stability.

This paper explores object detection in the small data regime, where only a limited number of annotated bounding boxes are available due to data rarity and annotation expense. This is a common challenge today with machine learning being applied to many new tasks where obtaining training data is more challenging, e.g. in medical images with rare diseases that doctors sometimes only see once in their life-time. In this work we explore this problem from a generative modeling perspective by learning to generate new images with associated bounding boxes, and using these for training an object detector. We show that simply training previously proposed generative models does not yield satisfactory performance due to them optimizing for image realism rather than object detection accuracy. To this end we develop a new model with a novel unrolling mechanism that jointly optimizes the generative model and a detector such that the generated images improve the performance of the detector. We show this method outperforms the state of the art on two challenging datasets, disease detection and small data pedestrian detection, improving the average precision on NIH Chest X-ray by a relative 20% and localization accuracy by a relative 50%., Comment: Published in ICCV 2019

Understanding the spatial relations between objects in images is a surprisingly challenging task. A chair may be "behind" a person even if it appears to the left of the person in the image (depending on which way the person is facing). Two students that appear close to each other in the image may not in fact be "next to" each other if there is a third student between them. We introduce SpatialSense, a dataset specializing in spatial relation recognition which captures a broad spectrum of such challenges, allowing for proper benchmarking of computer vision techniques. SpatialSense is constructed through adversarial crowdsourcing, in which human annotators are tasked with finding spatial relations that are difficult to predict using simple cues such as 2D spatial configuration or language priors. Adversarial crowdsourcing significantly reduces dataset bias and samples more interesting relations in the long tail compared to existing datasets. On SpatialSense, state-of-the-art recognition models perform comparably to simple baselines, suggesting that they rely on straightforward cues instead of fully reasoning about this complex task. The SpatialSense benchmark provides a path forward to advancing the spatial reasoning capabilities of computer vision systems. The dataset and code are available at https://github.com/princeton-vl/SpatialSense., Accepted to ICCV 2019

Ensemble learning aims to train and combine multiple base learners in the hope of improving the overall performance. Existing ensemble algorithms rely on a centralized framework where each base learner has access to the entire training dataset. We combine the technology of blockchains which is mainly used for data validation with ensemble learning and propose a decentralized framework where data are distributed among multiple base learners, who exchange their respective data to improve the collective predictive abilities. We develop two realizations of this framework, based on static and dynamic decision trees, respectively. We evaluate our methods over 20 real-world datasets and compare them against other centralized ensemble methods. Experimental results show that the proposed method obtains improved accuracy scores through sample exchange and achieves competitive performance with state-of-the-art ensemble methods whereas the base learners store only a small fraction of the samples.

Highly expressive models such as deep neural networks (DNNs) have been widely applied to various applications. However, recent studies show that DNNs are vulnerable to adversarial examples, which are carefully crafted inputs aiming to mislead the predictions. Currently, the majority of these studies have focused on perturbation added to image pixels, while such manipulation is not physically realistic. Some works have tried to overcome this limitation by attaching printable 2D patches or painting patterns onto surfaces, but can be potentially defended because 3D shape features are intact. In this paper, we propose meshAdv to generate "adversarial 3D meshes" from objects that have rich shape features but minimal textural variation. To manipulate the shape or texture of the objects, we make use of a differentiable renderer to compute accurate shading on the shape and propagate the gradient. Extensive experiments show that the generated 3D meshes are effective in attacking both classifiers and object detectors. We evaluate the attack under different viewpoints. In addition, we design a pipeline to perform black-box attack on a photorealistic renderer with unknown rendering parameters., Comment: Published in IEEE CVPR2019

INTRODUCTION AND OBJECTIVES:Deep learning is a subset of machine learning and artificial intelligence in which algorithms are developed to train machines/computers to detect a specific feature or c...

Polydactyly is a condition in which the supernumerary digits or duplications of digital parts manifest either as an integral part or in association with ∼300 well-defined syndromes. Polydactyly is the most frequently observed congenital limb anomaly at birth. The reported incidence of polydactyly

We introduce a new embedding model to represent movie characters and their interactions in a dialogue by encoding in the same representation the language used by these characters as well as information about the other participants in the dialogue. We evaluate the performance of these new character embeddings on two tasks: (1) character relatedness, using a dataset we introduce consisting of a dense character interaction matrix for 4,378 unique character pairs over 22 hours of dialogue from eighteen movies; and (2) character relation classification, for fine- and coarse-grained relations, as well as sentiment relations. Our experiments show that our model significantly outperforms the traditional Word2Vec continuous bag-of-words and skip-gram models, demonstrating the effectiveness of the character embeddings we introduce. We further show how these embeddings can be used in conjunction with a visual question answering system to improve over previous results.

Flapping wing deformation influences the aerodynamics of insect flight. This deformation is dictated by the dynamical properties of the insect wing, particularly its vibration spectra and mode shapes. However, researchers have not yet developed artificial insect wings with vibration spectra and mode shapes that are identical to their biological counterparts. The goal of the present work is to develop artificial insect wings that are both isospectral and isomodal with respect to real insect wings. To do so, we characterized hawkmoth Manduca sexta wings using experimental modal analyses. From these results, we created artificial wings using additive manufacturing and heat molding. Between artificial and real wings, the first two natural frequencies differ by 7% and 16% respectively, with differences of 16% and 131% in gains evaluated at those natural frequencies. Vibration modes are similar as well. This work provides a foundation for more advanced wing design moving forward.

This paper presents a novel AFM-based 3D nanofabrication process using ultrasonic vibration assisted nanomachining. A set of three dimensional nanostructures on polymethyl methacrylate (PMMA) samples are fabricated with the assistance of high frequency in-plane circular xy-vibration and ultrasonic tip-sample z-vibration. Two methods for fabricating 3D nanostructures were investigated in this study, which are layer-by-layer nanomachining and one pass nanomachining with the depth controlled by setpoint force. Critical parameters in the process are identified, including setpoint force, overlap percentage, amplitude of z vibration and machining speed. By regulating these process parameters, multi-level 3D nanostructures were fabricated by multi-layer machining in vector mode and raster scan mode. Using different setpoint forces for regulating feature depths, other nanostructures, such as convex and concave circles, were fabricated in raster scan mode from gray-scale bitmap pattern images. Under each mode, 3D nanostructure over microscale area can be fabricated in just a few minutes with sub-10 nm resolution in z direction.

Fine-grained recognition concerns categorization at sub-ordinate levels, where the distinction between object classes is highly local. Compared to basic level recognition, fine-grained categorization can be more challenging as there are in general less data and fewer discriminative features. This necessitates the use of a stronger prior for feature selection. In this work, we include humans in the loop to help computers select discriminative features. We introduce a novel online game called "Bubbles" that reveals discriminative features humans use. The player's goal is to identify the category of a heavily blurred image. During the game, the player can choose to reveal full details of circular regions ("bubbles"), with a certain penalty. With proper setup the game generates discriminative bubbles with assured quality. We next propose the "BubbleBank" representation that uses the human selected bubbles to improve machine recognition performance. Finally, we demonstrate how to extend BubbleBank to a view-invariant 3D representation. Experiments demonstrate that our approach yields large improvements over the previous state of the art on challenging benchmarks.

We have seen remarkable recent progress in computational visual recognition, producing systems that can classify objects into thousands of different categories with increasing accuracy. However, one question that has received relatively less attention is "what labels should recognition systems output?" This paper looks at the problem of predicting category labels that mimic how human observers would name objects. This goal is related to the concept of entry-level categories first introduced by psychologists in the 1970s and 1980s. We extend these seminal ideas to study human naming at large scale and to learn computational models for predicting entry-level categories. Practical applications of this work include improving human-focused computer vision applications such as automatically generating a natural language description for an image or text-based image search.

Background: Traditional Chinese medicine injections (TCMJ) used in the treatment of severe pneumonia have been widely implemented in clinical practice, but their overall efficacy and safety remain unclear. This paper aims to evaluate the efficacy and safety of TCMJ in the treatment of severe pneumonia. Methods: PubMed, EMbase, Cochrane Library, Web of Science, China National Knowledge Infrastructure, WanFang, and the Chongqing VIP Chinese Science and Technology Periodical Database were all searched for randomized controlled trials focusing on the administration of TCMJ for severe pneumonia. Two researchers independently screened titles, abstracts, full texts, and extracted relevant data. The RevMan 5.3 software (The Cochrane Collaboration, Software Update, Oxford, UK) and Stata 14 software (STATA Corporation, College Station, TX) were used for statistical analysis. Results: This study summarizes the related randomized controlled trials to assess the effect and safety of TCMJ in the treatment of severe pneumonia. Conclusion: This article provides theoretical support for the clinical application of TCMJ in the treatment of severe pneumonia. PROSPERO Registration number: CRD42020185072

Soil carbon dioxide (CO2) and nitrous oxide (N2O) emissions and carbon (C) sequestration are important processes for determining net greenhouse gas (GHG) emissions from maize fields. Farming management practices (FMPs) substantially affect crop productivity, soil organic carbon (SOC), and N2O emissions. However, relative knowledge is not sufficient, and further research is required for sustainable agriculture development. A multi-year experiment was performed on a spring maize field in the North China Plain (NCP) to quantify changes in SOC stocks, N2O emissions, and maize yields. Seven treatments were involved in the experiment, including no N fertilization (CK), farmers’ conventional N fertilization (FP), sole application of manure N (M), a combination of M and FP (MFP), FP combined with returned maize straw (SFP), fertilization with reduced synthetic N amount (RN), and fertilization with both reduced synthetic and manure N (RMN). The field measurements demonstrated that urea applications and returned maize straw increased N2O emission. The seasonal total N2O emissions under FP were significantly (P

Contemporary agriculture is shifting from a single-goal to a multi-goal strategy, which in turn requires choosing best management practice (BMP) based on an assessment of the biogeochemical effects of management alternatives. The bottleneck is the capacity of predicting the simultaneous effects of different management practice scenarios on multiple goals and choosing BMP among scenarios. The denitrification–decomposition (DNDC) model may provide an opportunity to solve this problem. We validated the DNDC model (version 95) using the observations of soil moisture and temperature, crop yields, aboveground biomass and fluxes of net ecosystem exchange of carbon dioxide, methane, nitrous oxide (N2O), nitric oxide (NO) and ammonia (NH3) from a wheat–maize cropping site in northern China. The model performed well for these variables. Then we used this model to simulate the effects of management practices on the goal variables of crop yields, NO emission, nitrate leaching, NH3 volatilization and net emission of greenhouse gases in the ecosystem (NEGE). Results showed that no-till and straw-incorporated practices had beneficial effects on crop yields and NEGE. Use of nitrification inhibitors decreased nitrate leaching and N2O and NO emissions, but they significantly increased NH3 volatilization. Irrigation based on crop demand significantly increased crop yield and decreased nitrate leaching and NH3 volatilization. Crop yields were hardly decreased if nitrogen dose was reduced by 15% or irrigation water amount was reduced by 25%. Two methods were used to identify BMP and resulted in the same BMP, which adopted the current crop cultivar, field operation schedules and full straw incorporation and applied nitrogen and irrigation water at 15 and 25% lower rates, respectively, than the current use. Our study indicates that the DNDC model can be used as a tool to assess biogeochemical effects of management alternatives and identify BMP.

We propose CornerNet, a new approach to object detection where we detect an object bounding box as a pair of keypoints, the top-left corner and the bottom-right corner, using a single convolution neural network. By detecting objects as paired keypoints, we eliminate the need for designing a set of anchor boxes commonly used in prior single-stage detectors. In addition to our novel formulation, we introduce corner pooling, a new type of pooling layer that helps the network better localize corners. Experiments show that CornerNet achieves a 42.2% AP on MS COCO, outperforming all existing one-stage detectors., Extended version with additional results. Test AP on MS COOO improved from 42.1% to 42.2% after a bug fix