Exploring Appropriate Reference Intervals and Clinical Decision Limits for Glucose-6-Phosphate Dehydrogenase Activity in Individuals From Guangzhou, China.

Background: Quantitative detection of glucose-6-phosphate dehydrogenase (G6PD) is commonly done to screen for G6PD deficiency. However, current reference intervals (RIs) of G6PD are unsuitable for evaluating G6PD-activity levels with local populations or associating G6PD variants with hemolysis risk to aid clinical decision-making. We explored appropriate RIs and clinical decision limits (CDLs) for G6PD activity in individuals from Guangzhou, China. Methods: We enrolled 5,852 unrelated individuals between 2020 and 2022 and screened their samples in quantitative assays for G6PD activity. We conducted further investigations, including G6PD genotyping, thalassemia genotyping, follow-up analysis, and statistical analysis, for different groups. Results: In Guangzhou, the RIs for the G6PD activities were 11.20-20.04 U/g Hb in male and 12.29-23.16 U/g Hb in female. The adjusted male median and normal male median (NMM) values were 15.47 U/g Hb and 15.51 U/g Hb, respectively. A threshold of 45% of the NMM could be used as a CDL to estimate the probability of G6PD variants. Our results revealed high hemolysis-risk CDLs (male: <10% of the NMM, female: <30% of the NMM), medium hemolysis-risk CDLs (male: 10%-45% of the NMM, female: 30%-79% of the NMM), and low hemolysis-risk CDLs (male: ≥ 45% of the NMM, female: ≥ 79% of the NMM). Conclusions: Collectively, our findings contribute to a more accurate evaluation of G6PD-activity levels within the local population and provide valuable insights for clinical decision-making. Specifically, identifying threshold values for G6PD variants and hemolysis risk enables improved prediction and management of G6PD deficiency, ultimately enhancing patient care and treatment outcomes.

Pulmonary Hypertension Induces Serotonin Hyperreactivity and Metabolic Reprogramming in Coronary Arteries via NOX1/4-TRPM2 Signaling Pathway.

BACKGROUND: Clinical evidence revealed abnormal prevalence of coronary artery (CA) disease in patients with pulmonary hypertension (PH). The mechanistic connection between PH and CA disease is unclear. Serotonin (5-hydroxytryptamine), reactive oxygen species, and Ca2+ signaling have been implicated in both PH and CA disease. Our recent study indicates that NOXs (NADPH [nicotinamide adenine dinucleotide phosphate] oxidases) and TRPM2 (transient receptor potential cation channel subfamily M member 2) are key components of their interplay. We hypothesize that activation of the NOX-TRPM2 pathway facilitates the remodeling of CA in PH. METHODS: Left and right CAs from chronic hypoxia and monocrotaline-induced PH rats were collected to study vascular reactivity, gene expression, metabolism, and mitochondrial function. Inhibitors or specific siRNA were used to examine the pathological functions of NOX1/4-TRPM2 in CA smooth muscle cells. RESULTS: Significant CA remodeling and 5-hydroxytryptamine hyperreactivity in the right CA were observed in PH rats. NOX1/4-mediated reactive oxygen species production coupled with TRPM2-mediated Ca2+ influx contributed to 5-hydroxytryptamine hyperresponsiveness. CA smooth muscle cells from chronic hypoxia-PH rats exhibited increased proliferation, migration, apoptosis, and metabolic reprogramming in an NOX1/4-TRPM2-dependent manner. Furthermore, the NOX1/4-TRPM2 pathway participated in mitochondrial dysfunction, involving mitochondrial DNA damage, reactive oxygen species production, elevated mitochondrial membrane potential, mitochondrial Ca2+ accumulation, and mitochondrial fission. In vivo knockdown of NOX1/4 alleviated PH and suppressed CA remodeling in chronic hypoxia rats. CONCLUSIONS: PH triggers an increase in 5-hydroxytryptamine reactivity in the right CA and provokes metabolic reprogramming and mitochondrial disruption in CA smooth muscle cells via NOX1/4-TRPM2 activation. This signaling pathway may play an important role in CA remodeling and CA disease in PH.

An improved YOLOv5s model for assessing apple graspability in automated harvesting scene.

Introduction: With continuously increasing labor costs, an urgent need for automated apple- Qpicking equipment has emerged in the agricultural sector. Prior to apple harvesting, it is imperative that the equipment not only accurately locates the apples, but also discerns the graspability of the fruit. While numerous studies on apple detection have been conducted, the challenges related to determining apple graspability remain unresolved. Methods: This study introduces a method for detecting multi-occluded apples based on an enhanced YOLOv5s model, with the aim of identifying the type of apple occlusion in complex orchard environments and determining apple graspability. Using bootstrap your own atent(BYOL) and knowledge transfer(KT) strategies, we effectively enhance the classification accuracy for multi-occluded apples while reducing data production costs. A selective kernel (SK) module is also incorporated, enabling the network model to more precisely identify various apple occlusion types. To evaluate the performance of our network model, we define three key metrics: APGA, APTUGA, and APUGA, representing the average detection accuracy for graspable, temporarily ungraspable, and ungraspable apples, respectively. Results: Experimental results indicate that the improved YOLOv5s model performs exceptionally well, achieving detection accuracies of 94.78%, 93.86%, and 94.98% for APGA, APTUGA, and APUGA, respectively. Discussion: Compared to current lightweight network models such as YOLOX-s and YOLOv7s, our proposed method demonstrates significant advantages across multiple evaluation metrics. In future research, we intend to integrate fruit posture and occlusion detection to f]urther enhance the visual perception capabilities of apple-picking equipment.

Intravascular Ultrasound-Based Fractional Flow Reserve for Predicting Prognosis after Percutaneous Coronary Intervention.

AccuFFRivus is an alternative to fractional flow reserve (FFR) based on intravascular ultrasound (IVUS) images for functional assessment of coronary stenosis. However, its prognostic impact in patients undergoing percutaneous coronary intervention (PCI) is still unclear. This retrospective study aimed to investigate the capability of AccuFFRivus in predicting prognosis. AccuFFRivus was calculated based on postoperative angiographic and IVUS images. Vessel-oriented clinical events (VOCE) at 2 years were recorded and analyzed. A total of 131 participants with 131 vessels were included in the study. VOCE occurred in 15 patients during 2-year follow-up. AccuFFRivus after PCI (post-AccuFFRivus) was significantly higher in the non-VOCE group than in the VOCE group (0.95 ± 0.03 vs. 0.91 ± 0.02, p < 0.001). Multivariate Cox regression showed that AccuFFRivus ≤ 0.94 was a strong independent predictor of VOCE during 2-year follow-up (hazard ratio 23.76, 95% confidence interval: 3.04-185.81, p < 0.001). The left panel displays the Receiver operating characteristics (ROC) curves of postoperative parameters (post-AccuFFRivus and post-MLA) versus vessel-oriented clinical events (VOCE) occurrence within 2-year follow-up. The right panel demonstrates Kaplan-Meier curves of VOCE stratified by the optimal cut-off of post-AccuFFRivus.

Estimation of Knee Joint Angle from Surface EMG Using Multiple Kernels Relevance Vector Regression.

In wearable robots, the application of surface electromyography (sEMG) signals in motion intention recognition is a hot research issue. To improve the viability of human-robot interactive perception and to reduce the complexity of the knee joint angle estimation model, this paper proposed an estimation model for knee joint angle based on the novel method of multiple kernel relevance vector regression (MKRVR) through offline learning. The root mean square error, mean absolute error, and R2_score are used as performance indicators. By comparing the estimation model of MKRVR and least squares support vector regression (LSSVR), the MKRVR performs better on the estimation of the knee joint angle. The results showed that the MKRVR can estimate the knee joint angle with a continuous global MAE of 3.27° ± 1.2°, RMSE of 4.81° ± 1.37°, and R2 of 0.8946 ± 0.07. Therefore, we concluded that the MKRVR for the estimation of the knee joint angle from sEMG is viable and could be used for motion analysis and the application of recognition of the wearer's motion intentions in human-robot collaboration control.

High-sensitivity vector bend sensor based on a fiber directional coupler inscribed by a femtosecond laser.

In this Letter, we demonstrate a high-sensitivity vector bend sensor based on a fiber directional coupler. The fiber directional coupler is composed of two parallel waveguides inscribed within a no-core fiber (NCF) by a femtosecond laser. Since the two written waveguides have closely matched refractive indices and geometries, the transmission spectrum of the fiber directional coupler possesses periodic resonant dips. Such a fiber directional coupler exhibits a good bending-dependent spectral shift response due to its asymmetric structure. Experimental results show that bending sensitivities of -97.11 nm/m-1 and 58.22 nm/m-1 are achieved for the 0° and 180° orientations in the curvature range of 0-0.62 m-1, respectively. In addition, the proposed fiber directional coupler is shown to be insensitive to external humidity changes, thus improving its suitability in high-accuracy bending measurements.

CircIFNGR2 enhances proliferation and migration of CRC and induces cetuximab resistance by indirectly targeting KRAS via sponging to MiR-30b.

Currently the clinical efficacy of colorectal cancer (CRC) which is the most common malignant tumors over the world has not reached an ideal level. Cetuximab, the monoclonal antibody targeting the extracellular domain of EGFR, has shown its great efficacy in the promotion of apoptosis and the inhibition of tumor cells-like characteristics in numerous cancers. However certain KRAS wild-type CRC patients unexpectedly show cetuximab resistance and the specific mechanism remains unclear. Circular RNAs (circRNAs) as the promising novel type of biomarkers in the cancer diagnosis and therapy, have been reported to be related with the drug resistance. In this study, with wondering the mechanism of cetuximab resistance in KRAS wild-type CRC patients, we evaluate the impact of circIFNGR2 on CRC and detect the association among circIFNGR2, miR-30b and KRAS via various experiments such as RT-qPCR, immunohistochemistry, luciferase assays, cell functional experiments and xenograft model. We conclude that circIFNGR2 induces cetuximab resistance in colorectal cancer cells by indirectly regulating target gene KRAS by sponging miR-30b at the post-transcriptional level. It is thus suggested that inhibition of circIFNGR2 can be a promising therapeutic strategy for malignant CRC patients with cetuximab resistance.

Augmentation Method for High Intra-Class Variation Data in Apple Detection.

Deep learning is widely used in modern orchard production for various inspection missions, which helps improve the efficiency of orchard operations. In the mission of visual detection during fruit picking, most current lightweight detection models are not yet effective enough to detect multi-type occlusion targets, severely affecting automated fruit-picking efficiency. This study addresses this problem by proposing the pioneering design of a multi-type occlusion apple dataset and an augmentation method of data balance. We divided apple occlusion into eight types and used the proposed method to balance the number of annotation boxes for multi-type occlusion apple targets. Finally, a validation experiment was carried out using five popular lightweight object detection models: yolox-s, yolov5-s, yolov4-s, yolov3-tiny, and efficidentdet-d0. The results show that, using the proposed augmentation method, the average detection precision of the five popular lightweight object detection models improved significantly. Specifically, the precision increased from 0.894 to 0.974, recall increased from 0.845 to 0.972, and mAP0.5 increased from 0.982 to 0.919 for yolox-s. This implies that the proposed augmentation method shows great potential for different fruit detection missions in future orchard applications.

An Unsupervised Monocular Visual Odometry Based on Multi-Scale Modeling.

Unsupervised deep learning methods have shown great success in jointly estimating camera pose and depth from monocular videos. However, previous methods mostly ignore the importance of multi-scale information, which is crucial for pose estimation and depth estimation, especially when the motion pattern is changed. This article proposes an unsupervised framework for monocular visual odometry (VO) that can model multi-scale information. The proposed method utilizes densely linked atrous convolutions to increase the receptive field size without losing image information, and adopts a non-local self-attention mechanism to effectively model the long-range dependency. Both of them can model objects of different scales in the image, thereby improving the accuracy of VO, especially in rotating scenes. Extensive experiments on the KITTI dataset have shown that our approach is competitive with other state-of-the-art unsupervised learning-based monocular methods and is comparable to supervised or model-based methods. In particular, we have achieved state-of-the-art results on rotation estimation.

Prediction of Differentially Expressed Genes and a Diagnostic Signature of Preeclampsia via Integrated Bioinformatics Analysis.

Background: Preeclampsia (PE), which has a high incidence rate worldwide, is a potentially dangerous syndrome to pregnant women and newborns. However, the exact mechanism of its pathogenesis is still unclear. In this study, we used bioinformatics analysis to identify hub genes, establish a logistic model, and study immune cell infiltration to clarify the physiopathogenesis of PE. Methods: We downloaded the GSE75010 and GSE10588 datasets from the GEO database and performed weighted gene coexpression network analysis (WGCNA) as well as Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. The online search tool for the retrieval of interacting genes and Cytoscape software were used to identify hub genes, which were then used to establish a logistic model. We also analyzed immune cell infiltration. Finally, we verified the expression of the genes included in the predictive model via RT-PCR. Results: A total of 100 and 212 differently expressed genes were identified in the GSE75010 and GSE10588 datasets, respectively, and after overlapping with WGCNA results, 17 genes were identified. KEGG and GO analyses further indicated the involvement of these genes in bioprocesses, such as gonadotropin secretion, immune cell infiltration, and the SMAD and MAPK pathways. Additionally, protein-protein interaction network analysis identified 10 hub genes, six (FLT1, FLNB, FSTL3, INHA, TREM1, and SLCO4A1) of which were used to establish a logistic model for PE. RT-PCR analysis also confirmed that, except FSTL3, these genes were upregulated in PE. Our results also indicated that macrophages played the most important role in immune cell infiltration in PE. Conclusion: This study identified 10 hub genes in PE and used 6 of them to establish a logistic model and also analyzed immune cell infiltration. These findings may enhance the understanding of PE and enable the identification of potential therapeutic targets for PE.

Deep-learning-based in-field citrus fruit detection and tracking.

Fruit yield estimation is crucial to establish fruit harvesting and marketing strategies. Recently, computer vision and deep learning techniques have been used to estimate citrus fruit yield and have exhibited a notable fruit detection ability. However, computer-vision-based citrus fruit counting has two key limitations: inconsistent fruit detection accuracy and double-counting of the same fruit. Using oranges as the experimental material, this paper proposes a deep-learning-based orange counting algorithm using video sequences to help overcome these problems. The algorithm consists of two sub-algorithms, OrangeYolo for fruit detection and OrangeSort for fruit tracking. The OrangeYolo backbone network is partially based on the YOLOv3 algorithm and improved upon to detect small object fruits at multiple scales. The network structure was adjusted to detect small-scale targets while enabling multiscale target detection. A channel attention and spatial attention multiscale fusion module was introduced to fuse the semantic features of the deep network with the shallow textural detail features. OrangeYolo can reach mean Average Precision (mAP) to 0.957 in the citrus dataset, which is higher than the 0.905, 0.911, and 0.917 that the YOLOv3, YOLOv4 and YOLOv5 algorithms. OrangeSort was designed to alleviate the double-counting problem of occluded fruits. A specific tracking region counting strategy and tracking algorithm based on motion displacement estimation are established. Six video sequences, which were taken from two fields containing 22 trees, were used as a validation dataset. The proposed method showed better performance (Mean Absolute Error(MAE) = 0.081, Standard Deviation(SD) = 0.08) compared to video-based manual counting and demonstrated more accurate results compared with existing standard Sort and DeepSort (MAE = 0.45, 1.212; SD = 0.4741, 1.3975; respectively).

SARS-CoV-2 causes a significant stress response mediated by small RNAs in the blood of COVID-19 patients.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has had a serious impact on the world. In this study, small RNAs from the blood of COVID-19 patients with moderate or severe symptoms were extracted for high-throughput sequencing and analysis. Interestingly, the levels of a special group of tRNA-derived small RNAs (tsRNAs) were found to be dramatically upregulated after SARS-CoV-2 infection, particularly in coronavirus disease 2019 (COVID-19) patients with severe symptoms. In particular, the 3'CCA tsRNAs from tRNA-Gly were highly consistent with the inflammation indicator C-reactive protein (CRP). In addition, we found that the majority of significantly changed microRNAs (miRNAs) were associated with endoplasmic reticulum (ER)/unfolded protein response (UPR) sensors, which may lead to the induction of proinflammatory cytokine and immune responses. This study found that SARS-CoV-2 infection caused significant changes in the levels of stress-associated small RNAs in patient blood and their potential functions. Our research revealed that the cells of COVID-19 patients undergo tremendous stress and respond, which can be reflected or regulated by small non-coding RNA (sncRNAs), thus providing potential thought for therapeutic intervention in COVID-19 by modulating small RNA levels or activities.

Lightweight Fruit-Detection Algorithm for Edge Computing Applications.

In recent years, deep-learning-based fruit-detection technology has exhibited excellent performance in modern horticulture research. However, deploying deep learning algorithms in real-time field applications is still challenging, owing to the relatively low image processing capability of edge devices. Such limitations are becoming a new bottleneck and hindering the utilization of AI algorithms in modern horticulture. In this paper, we propose a lightweight fruit-detection algorithm, specifically designed for edge devices. The algorithm is based on Light-CSPNet as the backbone network, an improved feature-extraction module, a down-sampling method, and a feature-fusion module, and it ensures real-time detection on edge devices while maintaining the fruit-detection accuracy. The proposed algorithm was tested on three edge devices: NVIDIA Jetson Xavier NX, NVIDIA Jetson TX2, and NVIDIA Jetson NANO. The experimental results show that the average detection precision of the proposed algorithm for orange, tomato, and apple datasets are 0.93, 0.847, and 0.850, respectively. Deploying the algorithm, the detection speed of NVIDIA Jetson Xavier NX reaches 21.3, 24.8, and 22.2 FPS, while that of NVIDIA Jetson TX2 reaches 13.9, 14.1, and 14.5 FPS and that of NVIDIA Jetson NANO reaches 6.3, 5.0, and 8.5 FPS for the three datasets. Additionally, the proposed algorithm provides a component add/remove function to flexibly adjust the model structure, considering the trade-off between the detection accuracy and speed in practical usage.

The ketogenic diet increases Neuregulin 1 expression via elevating histone acetylation and its anti-seizure effect requires ErbB4 kinase activity.

BACKGROUND: The ketogenic diet (KD)has been considered an effective treatment for epilepsy, whereas its underlying mechanisms remain obscure. We have previously reported that the KD feeding increased Neuregulin 1 (NRG1) expression in the hippocampus; disruption of NRG1 signaling by genetically deleting its receptor-ErbB4 abolished KD's effects on inhibitory synaptic activity and seizures. However, it is still unclear about the mechanisms underlying the effect of KD on NRG1 expression and whether the effects of KD require ErbB4 kinase activity. METHODS: The effects of the KD on NRG1 expression were assessed via western blotting and real-time PCR. Acetylation level at the Nrg1 promoter locus was examined using the chromatin immunoprecipitation technique. Kainic acid (KA)-induced acute seizure model was utilized to examine the effects of KD and histone deacetylase inhibitor-TSA on seizures. Synaptic activities in the hippocampus were recorded with the technique of electrophysiology. The obligatory role of ErbB4 kinase activity in KD's effects on seizures and inhibitory synaptic activity was evaluated by using ErbB kinase antagonist and transgenic mouse-T796G. RESULTS: We report that KD specifically increases Type I NRG1 expression in the hippocampus. Using the chromatin immunoprecipitation technique, we observe increased acetylated-histone occupancy at the Nrg1 promoter locus of KD-fed mice. Treatment of TSA dramatically elevates NRG1 expression and diminishes the difference between the effects of the control diet (CD) and KD. These data indicate that KD increases NRG1 expression via up-regulating histone acetylation. Moreover, both pharmacological and genetic inhibitions of ErbB4 kinase activity significantly block the KD's effects on inhibitory synaptic activity and seizure, suggesting an essential role of ErbB4 kinase activity. CONCLUSION: These results strengthen our understanding of the role of NRG1/ErbB4 signaling in KD and shed light on novel therapeutic interventions for epilepsy.

A chloroplast heat shock protein modulates growth and abiotic stress response in creeping bentgrass.

Small heat shock proteins (sHSPs), a family of the ubiquitous stress proteins in plants acting as molecular chaperones to protect other proteins from stress-induced damage, have been implicated in plant growth and development as well as plant response to environmental stress, especially heat stress. In this study, a chloroplast-localized sHSP, AsHSP26.8, was overexpressed in creeping bentgrass (Agrostis stolonifera L.) to study its role in regulating plant growth and stress response. Transgenic (TG) creeping bentgrass plants displayed arrested root development, slow growth rate, twisted leaf blades and are more susceptible to heat and salt but less sensitive to drought stress compared to wild-type (WT) controls. RNA-seq analysis revealed that AsHSP26.8 modulated the expression of genes in auxin signalling and stress-related genes such as those encoding HSPs, heat shock factors and other transcription factors. Our results provide new evidence demonstrating that AsHSP26.8 negatively regulates plant growth and development and plays differential roles in plant response to a plethora of diverse abiotic stresses.

Neuregulin 1/ErbB4 signaling contributes to the anti-epileptic effects of the ketogenic diet.

BACKGROUND: The ketogenic diet (KD) has been recognized as a potentially effective therapy to treat neuropsychiatric diseases, including epilepsy. Previous studies have indicated that KD treatment elevates γ-Amino butyric acid (GABA) levels in both human and murine brains, which presumably contributes to the KD's anti-seizure effects. However, this has not been systematically investigated at the synaptic level, and the underlying molecular mechanisms remain to be elucidated. METHODS: Kainic acid (KA)-induced acute and chronic seizure models were utilized to examine the effects of KD treatment on seizure threshold and epileptogenesis. Synaptic activities in the hippocampus were recorded with the technique of electrophysiology. The effects of the KD on Neuregulin 1 (Nrg1) expression were assessed via RNA sequencing, real-time PCR and Western blotting. The obligatory role of Nrg1 in KD's effects on seizures was evaluated through disruption of Nrg1 signaling in mice by genetically deleting its receptor-ErbB4. RESULTS: We found that KD treatment suppressed seizures in both acute and chronic seizure models and enhanced presynaptic GABA release probability in the hippocampus. By screening molecular targets linked to GABAergic activity with transcriptome analysis, we identified that KD treatment dramatically increased the Nrg1 gene expression in the hippocampus. Disruption of Nrg1 signaling by genetically deleting its receptor-ErbB4 abolished KD's effects on GABAergic activity and seizures. CONCLUSION: Our findings suggest a critical role of Nrg1/ErbB4 signaling in mediating KD's effects on GABAergic activity and seizures, shedding light on developing new therapeutic interventions to seizure control.

Natural variation of physiological traits, molecular markers, and chlorophyll catabolic genes associated with heat tolerance in perennial ryegrass accessions.

BACKGROUND: Identification of genetic diversity in heat tolerance and associated traits is of great importance for improving heat tolerance in cool-season grass species. The objectives of this study were to determine genetic variations in heat tolerance associated with phenotypic and physiological traits and to identify molecular markers associated with heat tolerance in a diverse collection of perennial ryegrass (Lolium perenne L.). RESULTS: Plants of 98 accessions were subjected to heat stress (35/30 °C, day/night) or optimal growth temperature (25/20 °C) for 24 d in growth chambers. Overall heat tolerance of those accessions was ranked by principal component analysis (PCA) based on eight phenotypic and physiological traits. Among these traits, electrolyte leakage (EL), chlorophyll content (Chl), relative water content (RWC) had high correlation coefficients (- 0.858, 0.769, and 0.764, respectively) with the PCA ranking of heat tolerance. We also found expression levels of four Chl catabolic genes (CCGs), including LpNYC1, LpNOL, LpSGR, and LpPPH, were significant higher in heat sensitive ryegrass accessions then heat tolerant ones under heat stress. Furthermore, 66 pairs of simple sequence repeat (SSR) markers were used to perform association analysis based on the PCA result. The population structure of ryegrass can be grouped into three clusters, and accessions in cluster C were relatively more heat tolerant than those in cluster A and B. SSR markers significantly associated with above-mentioned traits were identified (R2 > 0.05, p < 0.01)., including two pairs of markers located on chromosome 4 in association with Chl content and another four pairs of markers in association with EL. CONCLUSION: The result not only identified useful physiological parameters, including EL, Chl content, and RWC, and their associated SSR markers for heat-tolerance breeding of perennial ryegrass, but also highlighted the involvement of Chl catabolism in ryegrass heat tolerance. Such knowledge is of significance for heat-tolerance breeding and heat tolerance mechanisms in perennial ryegrass as well as in other cool-season grass species.

A NaI(Tl)-based radioactive noble-gas monitoring system used for radiation monitoring.

Radioactive noble-gas monitoring is necessary in nuclear facilities. A NaI(Tl)-based radioactive noble-gas monitoring system was developed. In order to increase the amount of air to be measured, the sample vessel of this system was larger than that of other systems, and was pressurized to about 5 × 105 Pa. In a laboratory experiment, technical ways to reduce the memory effect were investigated. In field tests, a method of spectra analysis was established and calibration coefficients and minimum detectable concentrations of 133Xe, 135Xe and 41Ar were calculated. Finally, detection ability was compared with other online monitoring systems.

Second-order Spectral Transform Block for 3D Shape Classification and Retrieval.

In this paper, we propose a novel network block, dubbed as second-order spectral transform block, for 3D shape retrieval and classification. This network block generalizes the second-order pooling to 3D surface by designing a learnable non-linear transform on the spectrum of the pooled descriptor. The proposed block consists of following two components. First, the second-order average (SO-Avr) and max-pooling (SOMax) operations are designed on 3D surface to aggregate local descriptors, which are shown to be more discriminative than the popular average-pooling or max-pooling. Second, a learnable spectral transform parameterized by mixture of power function is proposed to perform non-linear feature mapping in the space of pooled descriptors, i.e., manifold of symmetric positive definite matrix for SO-Avr, and space of symmetric matrix for SOMax. The proposed block can be plugged into existing network architectures to aggregate local shape descriptors for boosting their performance. We apply it to a shallow network for nonrigid 3D shape analysis and to existing networks for rigid shape analysis, where it improves the first-tier retrieval accuracy by 7.2% on SHREC'14 Real dataset and achieves state-of-the-art classification accuracy on ModelNet40. As an extension, we apply our block to 2D image classification, showing its superiority compared with traditional second-order pooling methods. We also provide theoretical and experimental analysis on stability of the proposed second-order spectral transform block.

ADMM-CSNet: A Deep Learning Approach for Image Compressive Sensing.

Compressive sensing (CS) is an effective technique for reconstructing image from a small amount of sampled data. It has been widely applied in medical imaging, remote sensing, image compression, etc. In this paper, we propose two versions of a novel deep learning architecture, dubbed as ADMM-CSNet, by combining the traditional model-based CS method and data-driven deep learning method for image reconstruction from sparsely sampled measurements. We first consider a generalized CS model for image reconstruction with undetermined regularizations in undetermined transform domains, and then two efficient solvers using Alternating Direction Method of Multipliers (ADMM) algorithm for optimizing the model are proposed. We further unroll and generalize the ADMM algorithm to be two deep architectures, in which all parameters of the CS model and the ADMM algorithm are discriminatively learned by end-to-end training. For both applications of fast CS complex-valued MR imaging and CS imaging of real-valued natural images, the proposed ADMM-CSNet achieved favorable reconstruction accuracy in fast computational speed compared with the traditional and the other deep learning methods.