Quality-Aware Selective Fusion Network for V-D-T Salient Object Detection.

Depth images and thermal images contain the spatial geometry information and surface temperature information, which can act as complementary information for the RGB modality. However, the quality of the depth and thermal images is often unreliable in some challenging scenarios, which will result in the performance degradation of the two-modal based salient object detection (SOD). Meanwhile, some researchers pay attention to the triple-modal SOD task, namely the visible-depth-thermal (VDT) SOD, where they attempt to explore the complementarity of the RGB image, the depth image, and the thermal image. However, existing triple-modal SOD methods fail to perceive the quality of depth maps and thermal images, which leads to performance degradation when dealing with scenes with low-quality depth and thermal images. Therefore, in this paper, we propose a quality-aware selective fusion network (QSF-Net) to conduct VDT salient object detection, which contains three subnets including the initial feature extraction subnet, the quality-aware region selection subnet, and the region-guided selective fusion subnet. Firstly, except for extracting features, the initial feature extraction subnet can generate a preliminary prediction map from each modality via a shrinkage pyramid architecture, which is equipped with the multi-scale fusion (MSF) module. Then, we design the weakly-supervised quality-aware region selection subnet to generate the quality-aware maps. Concretely, we first find the high-quality and low-quality regions by using the preliminary predictions, which further constitute the pseudo label that can be used to train this subnet. Finally, the region-guided selective fusion subnet purifies the initial features under the guidance of the quality-aware maps, and then fuses the triple-modal features and refines the edge details of prediction maps through the intra-modality and inter-modality attention (IIA) module and the edge refinement (ER) module, respectively. Extensive experiments are performed on VDT-2048 dataset, and the results show that our saliency model consistently outperforms 13 state-of-the-art methods with a large margin. Our code and results are available at https://github.com/Lx-Bao/QSFNet.

Composition of breast milk from mothers of premature and full-term infants and its influence in Z-Scores for infant physical growth.

BACKGROUND: Breast milk contains various crucial nutrients and biologically active substances and is ideal for newborns. This study aimed to analyze the composition of breast milk from mothers of premature and full-term infants and its influences on the growth of infants. METHODS: Infant-mother dyads examined at our Hospital (March 2016 to May 2017) were included. Milk was collected at 0-1 month, 2-3 months, and 5-6 months and analyzed using a MIRIS human milk analyzer. Z-scores of weight-for-length (WLZ), weight-for-age (WAZ), and length-for-age (LAZ) were calculated. RESULTS: This study included full-term (> 37 weeks of gestation, n = 177) and premature (< 37 weeks, n = 94) infant-mother dyads. The premature infants showed higher ΔWAZ, ΔLAZ, and ΔWLZ from infancy to toddlerhood for the physical growth speed, compared with term infants (P < 0.001). All proteins and true protein components of breast milk decreased with infants' age (P < 0.001). For premature and full-term infants, differences in ΔWAZ and ΔLAZ from birth to infancy and the difference in ΔLAZ, WAZ, and LAZ in toddlerhood were positively associated with non-protein nitrogen (NPN) (all P < 0.05), while the Z-score differences in ΔWLZ from birth to infancy were negatively associated with NPN (all P < 0.05). For premature babies, from birth to infancy stage, ΔWAZ was positively correlated with NPN and carbohydrates while negatively correlated with dry matter (all P < 0.05), and ΔLAZ correlated with NPN (ß = 0.428, P = 0.005). CONCLUSION: Breastfeeding helped premature infants compensatory growth when compared to term infants. Whileduring early infancy stage ΔWLZ gain was negatively associated with increased amounts of NPN in breast milk. This might mean although NPN increase the Z-scores of weight-for-age and length-for-age, with no rise in adipose tissue mass.

Multisystem inflammatory syndrome in children with Kawasaki disease-like manifestations in MMA.

Not required for Clinical Vignette.

Lightweight Cross-Modal Information Mutual Reinforcement Network for RGB-T Salient Object Detection.

RGB-T salient object detection (SOD) has made significant progress in recent years. However, most existing works are based on heavy models, which are not applicable to mobile devices. Additionally, there is still room for improvement in the design of cross-modal feature fusion and cross-level feature fusion. To address these issues, we propose a lightweight cross-modal information mutual reinforcement network for RGB-T SOD. Our network consists of a lightweight encoder, the cross-modal information mutual reinforcement (CMIMR) module, and the semantic-information-guided fusion (SIGF) module. To reduce the computational cost and the number of parameters, we employ the lightweight module in both the encoder and decoder. Furthermore, to fuse the complementary information between two-modal features, we design the CMIMR module to enhance the two-modal features. This module effectively refines the two-modal features by absorbing previous-level semantic information and inter-modal complementary information. In addition, to fuse the cross-level feature and detect multiscale salient objects, we design the SIGF module, which effectively suppresses the background noisy information in low-level features and extracts multiscale information. We conduct extensive experiments on three RGB-T datasets, and our method achieves competitive performance compared to the other 15 state-of-the-art methods.

Diagnosis of DC series arc fault based on multiple wavelet transform and optimal fractional wavelet energy entropy.

In the DC distribution system, the propagation of arc noise can interfere with normal lines, and accurate and timely diagnosis of the location of series arc fault (SAF) is a challenging problem. In this article, a SAF diagnosis method is proposed from a system perspective, which can accurately identify the fault line. First, multiple wavelet transform is used to decompose the currents of different lines, and the fractional wavelet energy entropy is extracted to construct the feature vector. Then, random forest is employed to analyze the importance of features and to select the optimal features. Finally, a kernel extreme learning machine can fuse the features and output the diagnosis results. The offline experimental results indicate that the proposed method has a diagnosis accuracy of 99.82%, which is higher than those of nine comparison methods, and the effectiveness and advancement of the proposed method are verified. The online experimental results show that the proposed method can diagnose SAF within 110 ms, and the diagnosis speed is able to satisfy the requirements of UL1699B. Moreover, under transient conditions, the proposed method can effectively avoid false alarms and maintain stability.

Edge-Guided Recurrent Positioning Network for Salient Object Detection in Optical Remote Sensing Images.

Optical remote sensing images (RSIs) have been widely used in many applications, and one of the interesting issues about optical RSIs is the salient object detection (SOD). However, due to diverse object types, various object scales, numerous object orientations, and cluttered backgrounds in optical RSIs, the performance of the existing SOD models often degrade largely. Meanwhile, cutting-edge SOD models targeting optical RSIs typically focus on suppressing cluttered backgrounds, while they neglect the importance of edge information which is crucial for obtaining precise saliency maps. To address this dilemma, this article proposes an edge-guided recurrent positioning network (ERPNet) to pop-out salient objects in optical RSIs, where the key point lies in the edge-aware position attention unit (EPAU). First, the encoder is used to give salient objects a good representation, that is, multilevel deep features, which are then delivered into two parallel decoders, including: 1) an edge extraction part and 2) a feature fusion part. The edge extraction module and the encoder form a U-shape architecture, which not only provides accurate salient edge clues but also ensures the integrality of edge information by extra deploying the intraconnection. That is to say, edge features can be generated and reinforced by incorporating object features from the encoder. Meanwhile, each decoding step of the feature fusion module provides the position attention about salient objects, where position cues are sharpened by the effective edge information and are used to recurrently calibrate the misaligned decoding process. After that, we can obtain the final saliency map by fusing all position attention cues. Extensive experiments are conducted on two public optical RSIs datasets, and the results show that the proposed ERPNet can accurately and completely pop-out salient objects, which consistently outperforms the state-of-the-art SOD models.

Deconstruction of the Optimal Design of Urban Road Interchange Based on the Integration of Smart Transportation and Big Data.

Urban interchange is the core hub connecting various regions, and it is of great significance for alleviating the problem of traffic congestion. In the process of urban interchange design, it is impossible to strictly control the traffic volume, interchange types, and standards by relying on traditional technologies. Smart transportation and big data are emerging technologies based on data, which can provide technical support for design and decision making. Based on this, this paper first uses smart transportation and big data technology to predict the traffic volume of Nancheng New District, so as to calculate the future development trend of the target area. Then, on the basis of traffic volume, the article uses smart transportation and big data technology to optimize the original urban interchange design scheme from the aspects of traffic capacity, safety, economic benefits, and environmental benefits. Finally, the article evaluates the optimized urban interchange scheme by means of comprehensive quantitative indicators and evaluation methods. Experiments show that the traffic capacity of the interchange on the outer ring road optimized by smart transportation and big data has increased to 72.6%, and the environmental coordination has increased from 45.2% to 55.2%. Moreover, the design aesthetics of the urban interchange after optimized design based on smart transportation and big data has increased to 65.9%. In addition, the comprehensive evaluation value of the urban interchange after optimization of smart transportation and big data reached 82.6. This fully shows that the optimal design of urban interchange based on the integration of smart transportation and big data can greatly improve the traffic capacity of urban roads.

Regional Brain Fusion: Graph Convolutional Network for Alzheimer's Disease Prediction and Analysis.

Alzheimer's disease (AD) has raised extensive concern in healthcare and academia as one of the most prevalent health threats to the elderly. Due to the irreversible nature of AD, early and accurate diagnoses are significant for effective prevention and treatment. However, diverse clinical symptoms and limited neuroimaging accuracy make diagnoses challenging. In this article, we built a brain network for each subject, which assembles several commonly used neuroimaging data simply and reasonably, including structural magnetic resonance imaging (MRI), diffusion-weighted imaging (DWI), and amyloid positron emission tomography (PET). Based on some existing research results, we applied statistical methods to analyze (i) the distinct affinity of AD burden on each brain region, (ii) the topological lateralization between left and right hemispheric sub-networks, and (iii) the asymmetry of the AD attacks on the left and right hemispheres. In the light of advances in graph convolutional networks for graph classifications and summarized characteristics of brain networks and AD pathologies, we proposed a regional brain fusion-graph convolutional network (RBF-GCN), which is constructed with an RBF framework mainly, including three sub-modules, namely, hemispheric network generation module, multichannel GCN module, and feature fusion module. In the multichannel GCN module, the improved GCN by our proposed adaptive native node attribute (ANNA) unit embeds within each channel independently. We not only fully verified the effectiveness of the RBF framework and ANNA unit but also achieved competitive results in multiple sets of AD stages' classification tasks using hundreds of experiments over the ADNI clinical dataset.

Inter-alpha Inhibitor Proteins Ameliorate Brain Injury and Improve Behavioral Outcomes in a Sex-Dependent Manner After Exposure to Neonatal Hypoxia Ischemia in Newborn and Young Adult Rats.

Hypoxic-ischemic (HI) brain injury is a major contributor to neurodevelopmental morbidities. Inter-alpha inhibitor proteins (IAIPs) have neuroprotective effects on HI-related brain injury in neonatal rats. However, the effects of treatment with IAIPs on sequential behavioral, MRI, and histopathological abnormalities in the young adult brain after treatment with IAIPs in neonates remain to be determined. The objective of this study was to examine the neuroprotective effects of IAIPs at different neurodevelopmental stages from newborn to young adults after exposure of neonates to HI injury. IAIPs were given as 11-sequential 30-mg/kg doses to postnatal (P) day 7-21 rats after right common carotid artery ligation and exposure to 90 min of 8% oxygen. The resulting brain edema and injury were examined by T2-weighted magnetic resonance imaging (MRI) and cresyl violet staining, respectively. The mean T2 values of the ipsilateral hemisphere from MRI slices 6 to 10 were reduced in IAIP-treated HI males + females on P8, P9, and P10 and females on P8, P9, P10, and P14. IAIP treatment reduced hemispheric volume atrophy by 44.5 ± 29.7% in adult male + female P42 rats and improved general locomotor abilities measured by the righting reflex over time at P7.5, P8, and P9 in males + females and males and muscle strength/endurance measured by wire hang on P16 in males + females and females. IAIPs provided beneficial effects during the learning phase of the Morris water maze with females exhibiting beneficial effects. IAIPs confer neuroprotection from HI-related brain injury in neonates and even in adult rats and beneficial MRI and behavioral benefits in a sex-dependent manner.

Visible-light driven photocatalytic performance of eco-friendly cobalt-doped ZnO nanoarrays: Influence of morphology, cobalt doping, and photocatalytic efficiency.

In order to remove the organic pollution from water environment, Co2+-doped ZnO nanoarray photocatalyst was prepared through a hydrothermal process. The influences of Co2+ doping amount and hydrothermal temperature on the nanostructure and photocatalytic performance of Co2+-doped ZnO nanoarray were discussed in detail. The standard ZnO structure and nanoarray morphology of Co2+-doped ZnO samples were achieved and the absorption of visible light was also realized through Co2+ doping. The 2% Co2+-doped ZnO nanoarray prepared at 95 °C exhibited excellent photocatalytic activity and could degrade 96% of methylene blue solution within 120 min under visible light. Furthermore, the as-prepared 2% Co2+-doped ZnO nanoarray still maintained 91% for removal rate after 3 cycles of photocatalytic degradation, showed good photocatalytic activity and recyclability. All results indicate that ZnO nanoarray with Co2+ doping has a potential application in visible light photocatalysis for environmental protection and pollution control.

Structure and luminescence investigation of Gd3+-sensitized perovskite CaLa4Ti4O15:Eu3+: A novel red-emitting phosphor for high-performance white light-emitting diodes and plants lighting.

A novel perovskite CaLa4Ti4O15:Eu3+ red-emitting phosphor was synthesized via a sol-combustion method, and Gd3+ was further co-doped into structure to improve the luminescence performance. The effects of Eu3+ doping and Gd3+ co-doping concentrations on the microstructure and luminescence properties were investigated. The red emission peaks of as-prepared phosphors originate from the 5D0→7Fj electron transitions of Eu3+ ions. Under 273 nm excitation, the luminescence intensity of Eu3+ was significantly enhanced through the energy transfer between Gd3+ and Eu3+ in CaLa4Ti4O15, and the luminescence intensity was also improved even under the excitation of 394 nm. By combining red-emitting CaLa4Ti4O15:Eu3+, Gd3+ phosphor with commercial blue and green phosphors on n-UV chip (λ = 395 nm), an eye-friendly w-LEDs with appropriate correlated color temperature (4761 K) and high color rendering index (Ra = 93.1) has been realized. The electroluminescence spectrum of the packaged red LED have an excellent match with the PR absorption of plants. In addition, when introducing CaLa4Ti4O15:Eu3+, Gd3+ phosphor into a commercial w-LED with YAG:Ce3+, the adjustable chromaticity parameters like CCT and CRI values can be obtained. These results demonstrated that the as-prepared CaLa4Ti4O15:Eu3+, Gd3+ phosphor is an outstanding candidate as the red component for the application of w-LEDs and plants lighting.

Learning Frequency Domain Priors for Image Demoireing.

Image demoireing is a multi-faceted image restoration task involving both moire pattern removal and color restoration. In this paper, we raise a general degradation model to describe an image contaminated by moire patterns, and propose a novel multi-scale bandpass convolutional neural network (MBCNN) for single image demoireing. For moire pattern removal, we propose a multi-block-size learnable bandpass filters (M-LBFs), based on a block-wise frequency domain transform, to learn the frequency domain priors of moire patterns. We also introduce a new loss function named Dilated Advanced Sobel loss (D-ASL) to better sense the frequency information. For color restoration, we propose a two-step tone mapping strategy, which first applies a global tone mapping to correct for a global color shift, and then performs local fine tuning of the color per pixel. To determine the most appropriate frequency domain transform, we investigate several transforms including DCT, DFT, DWT, learnable non-linear transform and learnable orthogonal transform. We finally adopt the DCT. Our basic model won the AIM2019 demoireing challenge. Experimental results on three public datasets show that our method outperforms state-of-the-art methods by a large margin.

Dynamic Selective Network for RGB-D Salient Object Detection.

RGB-D saliency detection is receiving more and more attention in recent years. There are many efforts have been devoted to this area, where most of them try to integrate the multi-modal information, i.e. RGB images and depth maps, via various fusion strategies. However, some of them ignore the inherent difference between the two modalities, which leads to the performance degradation when handling some challenging scenes. Therefore, in this paper, we propose a novel RGB-D saliency model, namely Dynamic Selective Network (DSNet), to perform salient object detection (SOD) in RGB-D images by taking full advantage of the complementarity between the two modalities. Specifically, we first deploy a cross-modal global context module (CGCM) to acquire the high-level semantic information, which can be used to roughly locate salient objects. Then, we design a dynamic selective module (DSM) to dynamically mine the cross-modal complementary information between RGB images and depth maps, and to further optimize the multi-level and multi-scale information by executing the gated and pooling based selection, respectively. Moreover, we conduct the boundary refinement to obtain high-quality saliency maps with clear boundary details. Extensive experiments on eight public RGB-D datasets show that the proposed DSNet achieves a competitive and excellent performance against the current 17 state-of-the-art RGB-D SOD models.

Comparison of 2 dynamic conformal arc plans based on high-dose rate flattening filter free beams for peripheral lung cancer. / åŸºäºŽå‘¨å›´åž‹è‚ºç™Œé«˜å‰‚é‡çŽ‡æ— å‡æ•´æ»¤è¿‡å™¨æ¨¡å¼ä¸‹çš„2种动态适形弧计划比较.

OBJECTIVES: To compare 2 dynamic conformal arc plans based on the high dose rate flattening filter free (FFF) beams, and to evaluate the dosimetric differences. METHODS: A total of 20 patients with early peripheral non-small cell lung cancer were selected, and 2 dynamic conformal arc plans were designed in the Eclipse 10.0 treatment planning system (TPS). One of them was based on tumor-center (T-DCA), and the other was based on iso-center (Iso-DCA). Both plans were created by using the Truebeam linear accelerator, based on 6 MV FFF photons with a dose rate at 1 400 monitor unit (MU)/min. All patients received the prescribed dose of 4 800 cGy in 4 fractions (1 200 cGy/fraction). Target coverage and organ at risk limits were planned and designed according to the Radiation Therapy Oncology Group (RTOG) Criteria, and were compared between the T-DCA and the Iso-DCA plans. RESULTS: There was no significant difference in the target coverage between the T-DCA and Iso-DCA plans (P>0.05). Conformal index and homogeneity index had no significant differences (both P>0.05), but the percentage of the maximum dose in any direction 2 cm away from the planned target area (D2 cm) and the ratio of the volume wrapped by the isodose line of 50% prescription dose to the volume of the planned target area (R50%) showed significant differences (both P<0.05). The MU of the Iso-DCA plan was increased by 21% compared with that of the T-DCA plan. Except the maximum dose of spinal cord and esophagus, there was no significant difference in the other dosimetric parameters of the organs at risk between the T-DCA and the Iso-DCA plans (all P>0.05). CONCLUSIONS: The dose fall-off of Iso-DCA plan is better than T-DCA plan, but the T-DCA plan is consistently superior in sparing dose to spinal cord and esophagus, and the T-DCA plan has fewer MU.

Preclinical Characterization of the Selective JAK1 Inhibitor LW402 for Treatment of Rheumatoid Arthritis.

BACKGROUND: Research on JAK family members as therapeutic targets for autoimmune diseases has brought tofacitinib and baricitinib into clinical for the treatment of rheumatoid arthritis and other autoimmune diseases. Despite the potent efficacy of these first-generation JAK inhibitors, their broad-spectrum JAK inhibition and adverse events warrant development of a JAK1-specific inhibitor to improve their safety profile. METHODS: In this study, we characterized a JAK1-specific inhibitor, LW402, on biochemical and human whole-blood assays. We further evaluated the therapeutic efficacy of LW402 in a rat adjuvant-induced arthritis (rAIA) model and a mouse collagen-induced arthritis (mCIA) model. The safety of LW402 was evaluated in both SpragueDawley rats and cynomolgus monkeys. RESULTS: LW402 exhibited potent nanomolar activity against JAK1 and showed a 45-fold selectivity for inhibition of JAK1- over JAK2-dependent signaling induced by either IL6 or GM-CSF in human whole-blood assays. In the rAIA model, oral dosing of LW402 resulted in a dose-dependent improvement in disease symptoms, including reduction in paw swelling, marked reduction in the inflammatory-cell infiltration to synovial tissue, and protection of articular cartilage and bone from damage. The therapeutic efficacy of LW402 correlated well with the plasma exposure of LW402 and the extent of pSTAT3 inhibition in white blood cells. LW402 also effectively eased disease symptoms in the mCIA model. Toxicity studies in the Sprague Dawley rats and cynomolgus monkeys established a ≥5x therapeutic window for LW402 as drug exposures of toxicity study NOAEL dose and pharmacology study ED50 dose were compared. CONCLUSION: We developed a novel JAK1-specific inhibitor LW402 with potent efficacy in rAIA and mCIA models. We established a good safety profile for LW402 in toxicity studies, and the overall superiority of LW402 should translated well to the clinical setting for the treatment of RA and other autoimmune diseases.

Tailoring PTV expansion to improve the dosimetry of post modified radical mastectomy intensity-modulated radiotherapy for left-sided breast cancer patients by using 4D CT combined with cone beam CT.

PURPOSE: Our study aimed to improve the dosimetry of post modified radical mastectomy intensity-modulated radiotherapy (PMRM-IMRT) for left-sided breast cancer patients by tailoring and minimizing PTV expansion three-dimensionally utilizing 4D CT combined with on-board cone beam CT (CBCT). METHODS: We enrolled a total of 10 consecutive left-sided breast cancer patients to undergo PMRM-IMRT. We measured the intra-fractional CTV displacement attributed to respiratory movement by defining 9 points on the left chest wall and quantifying their displacement by using the 4D CT, and measured the inter-fractional CTV displacement resulting from the integrated effect of respiratory movement, thoracic deformation and set up errors by using CBCT. We created 3 different PMRM-IMRT plans for each of the patients using PTVt (tailored PTV expansion three-dimensionally), PTV0.5 and PTV0.7 (isotropic 0.5- cm and isotropic 0.7- cm expanding margin of CTV), respectively. We performed paired samples t test to establish a hierarchy in terms of plan quality and dosimetric benefits. P < 0.05 was considered statistically significant. RESULTS: The inter-fractional CTV displacement (2.6 ± 2.2 mm vertically, 2.8 ± 2.3 mm longitudinally, and 1.7 ± 1.2 mm laterally) measured by CBCT was much larger than the intra-fractional one (0.5 ± 0.5 mm vertically, 0.5 ± 1.0 mm longitudinally, and 0.3 ± 0.3 mm laterally, respectively) measured by 4D CT. Intensity-modulated radiotherapy with tailored PTV expansion based on inter-fractional CTV displacement had dosimetrical advantages over those with PTV0.5 or those with PTV0.7 owing to its perfect PTV dose coverage and better OARs sparing(especially of heart and left lung). CONCLUSION: The CTV displacement in PMRM-IMRT predominantly arises from inter-fraction rather than from intra-fraction during natural respiration and differs in 3 coordinate axes either inter-fractionally or intra-fractionally. Tailoring and minimizing PTV expansion three-dimensionally significantly improves the dosimetry of PMRM-IMRT for left-sided breast cancer patients.

Hyaluronate-Based Self-Stabilized Nanoparticles for Immunosuppression Reversion and Immunochemotherapy in Osteosarcoma Treatment.

Immunotherapy is regarded as a potential strategy to combat cancer, especially when immunotherapy is combined with appropriate chemotherapy. However, the immunosuppressive tumor microenvironment (TME) and serious side effects extremely limit the application of immunotherapy. Herein, a self-stabilized hyaluronic acid nanoparticle is synthesized for tumor-targeted delivery of doxorubicin (DOX), cisplatin (CDDP), and resiquimod (R848) in osteosarcoma immunochemotherapy, which is referred to as CDDPNPDOX&R848. CDDPNPDOX&R848 exhibits sufficient stability, great pH responsibility, and brilliant tumor-targeting accumulation in vivo, which make it suitable for further in vivo applications. After intravenous injection, CDDPNPDOX&R848 can release the loaded cargoes under the acidic TME continuously. DOX can induce tumor cell apoptosis in combination with CDDP and trigger immunogenic cell death. More importantly, the immune-activated TME created by R848 can facilitate tumor-associated antigen presentation and antitumor immunity elicitation. Benefiting from the synergistic effect of chemotherapy and immunotherapy, the growth of tumors and lung metastasis was greatly inhibited by CDDPNPDOX&R848 in the K7M2 orthotopic osteosarcoma mouse model. Thus, this intelligent codelivery platform might be a competitive candidate for osteosarcoma immunochemotherapy.

The role of peroxisome proliferator-activated receptor gamma and adiponectin in children with Kawasaki disease.

OBJECTIVE: To investigate the relationship between peroxisome proliferator-activated receptor gamma (PPARγ) mRNA, serum adiponectin (ADP) and lipids in paediatric patients with Kawasaki disease (KD). METHODS: This prospective study enrolled paediatric patients with KD and grouped them according to the presence or absence of coronary artery lesions (CAL). A group of healthy age-matched children were recruited as the control group. The levels of PPARγ mRNA, serum ADP and lipids were compared between the groups. Receiver operating characteristic (ROC) curve analysis was undertaken to determine if the PPARγ mRNA level could be used as a predictive biomarker of CAL prognosis. RESULTS: The study enrolled 42 patients with KD (18 with CAL [CAL group] and 24 without CAL [NCAL group]) and 20 age-matched controls. PPARγ mRNA levels in patients with KD were significantly higher than those in the controls; but significantly lower in the CAL group than the NCAL group. ROC curve analysis demonstrated that the PPARγ mRNA level provided good predictive accuracy for the prognosis of CAL. There was no association between PPARγ, ADP and lipid levels. CONCLUSION: There was dyslipidaemia in children with KD, but there was no correlation with PPARγ and ADP. PPARγ may be a predictor of CAL in patients with KD with good predictive accuracy.

A Robust Lane Detection Model Using Vertical Spatial Features and Contextual Driving Information.

The quality of detected lane lines has a great influence on the driving decisions of unmanned vehicles. However, during the process of unmanned vehicle driving, the changes in the driving scene cause much trouble for lane detection algorithms. The unclear and occluded lane lines cannot be clearly detected by most existing lane detection models in many complex driving scenes, such as crowded scene, poor light condition, etc. In view of this, we propose a robust lane detection model using vertical spatial features and contextual driving information in complex driving scenes. The more effective use of contextual information and vertical spatial features enables the proposed model more robust detect unclear and occluded lane lines by two designed blocks: feature merging block and information exchange block. The feature merging block can provide increased contextual information to pass to the subsequent network, which enables the network to learn more feature details to help detect unclear lane lines. The information exchange block is a novel block that combines the advantages of spatial convolution and dilated convolution to enhance the process of information transfer between pixels. The addition of spatial information allows the network to better detect occluded lane lines. Experimental results show that our proposed model can detect lane lines more robustly and precisely than state-of-the-art models in a variety of complex driving scenarios.

An integrated classification model for incremental learning.

Incremental Learning is a particular form of machine learning that enables a model to be modified incrementally, when new data becomes available. In this way, the model can adapt to the new data without the lengthy and time-consuming process required for complete model re-training. However, existing incremental learning methods face two significant problems: 1) noise in the classification sample data, 2) poor accuracy of modern classification algorithms when applied to modern classification problems. In order to deal with these issues, this paper proposes an integrated classification model, known as a Pre-trained Truncated Gradient Confidence-weighted (Pt-TGCW) model. Since the pre-trained model can extract and transform image information into a feature vector, the integrated model also shows its advantages in the field of image classification. Experimental results on ten datasets demonstrate that the proposed method outperform the original counterparts.