Search

In real-world scenarios, it is common that a video contains multiple actors and their activities. Selectively localizing one specific actor and its action spatially and temporally via a language query becomes a vital and challenging task. Existing fully supervised methods require extensive elaborately annotated data and are sensitive to the class labels, which cannot satisfy real-world applications’ needs. Thus, we introduce the task of weakly supervised actor-action video segmentation from a sentence query (AAVSS) in this work, where only the video-sentence pairs are provided. To the best of our knowledge, our work is the first to perform AAVSS under weakly supervised situations. However, this task is extremely challenging not only because the task aims to learn the complex interactions between two heterogeneous modalities but also because the task needs to learn fine-grained analysis of video content without pixel-level annotations. To overcome the challenges, we propose a two-stage network. The network first follows the sentence guidance to localize the candidate region and then performs segmentation to achieve selective segmentation. Specifically, a novel tracker-based clip-level multiple instance learning paradigm is proposed in this article to learn the matches between regions and sentences, which makes our two-stage network robust to the region proposal network. Furthermore, two intrinsic characteristics of the video, temporal consistency and motion information, are utilized in companion with the weak supervision to facilitate the region-query matching. Through extensive experiments, the proposed method achieves comparable performance to state-of-the-art fully supervised approaches on two large-scale benchmarks, including A2D Sentences and J-HMDB Sentences.

Recently, we have observed an exponential increase of user-generated content (UGC) videos. The distinguished characteristic of UGC videos originates from the video production and delivery chain, as they are usually acquired and processed by non-professional users before uploading to the hosting platforms for sharing. As such, these videos usually undergo multiple distortion stages that may affect visual quality before ultimately being viewed. Inspired by the increasing consensus that the optimization of the video coding and processing shall be fully driven by the perceptual quality, in this paper, we propose to study the quality of the UGC videos from both objective and subjective perspectives. We first construct a UGC video quality assessment (VQA) database, aiming to provide useful guidance for the UGC video coding and processing in the hosting platform. The database contains source UGC videos uploaded to the platform and their transcoded versions that are ultimately enjoyed by end-users, along with their subjective scores. Furthermore, we develop an objective quality assessment algorithm that automatically evaluates the quality of the transcoded videos based on the corrupted reference, which is in accordance with the application scenarios of UGC video sharing in the hosting platforms. The information from the corrupted reference is well leveraged and the quality is predicted based on the inferred quality maps with deep neural networks (DNN). Experimental results show that the proposed method yields superior performance. Both subjective and objective evaluations of the UGC videos also shed lights on the design of perceptual UGC video coding.

Recent achievements in large-scale pre-trained models like GPT-3 and PanGu-α showed amazing performances in many downstream tasks, which makes AI friendlier toward industrial users. Deep learning has been recognized as the most promising technology for pharmaceuticals, a powerful molecule pre-trained model could save researchers tons of time. In chemistry classes, the students learn two molecule representations, the molecular formula and the structure formula, and learn to translate them from one way to the other. Inspired by this, we developed a novel deep learning architecture using a graph-to-sequence asymmetric conditional variational autoencoders, called PanGu Drug Model, which can appropriately characterize the molecule from both representations and improve the performance of downstream drug discovery tasks. After pretrained with 1.7 billion small molecules, our proposed model achieved the state-of-the-art results in 20 drug discovery tasks, such as molecule property prediction (predict ADMET properties, compound-target interactions, drug-drug interactions and chemical reaction productivity), molecule generation and molecule optimization. A new drug screening library of 100 million drug-like small molecules with 99.68% novelty was generated by PanGu molecule generator, which could efficiently produce novel compounds with similar physiochemical properties to given distribution, this library could be used to supplement existing compound databases. In addition, PanGu molecule optimizer could optimize the chemical structures of starting molecule with improved molecular property of interest. An automatic multi-objective optimization web application implemented by PanGu Drug Model is available at http://www.pangu-drug.com/.

Polyoxymethylene dimethyl ethers (PODE) are newly appeared promising oxygenated additive which can significantly reduce soot emissions of diesel engines. In this study, the spray and combustion characteristics of the blends of diesel with different PODE blending ratio of 0%, 20%, 50% and 100% by volume are investigated in a constant volume chamber (CVC) under 100 MPa, 120 MPa and 140 MPa injection pressures. The spray results show that liquid spray tip penetration (STP), average spray angle (ASA) and spray projected area (SPA) all decrease with PODE blending ratio under 4 MPa ambient pressure and 573 K ambient temperature. The addition of PODE in diesel can significantly reduce the soot formation and the integrated natural flame luminosity (INFL). Under 100 MPa injection pressure, the time integral sum KL in whole combustion duration of D100, P20, P50 and P100 are namely 107.35, 2.66, 0.27 and 0.08. When the PODE blending ratio are 50% and 100%, nearly no yellow flame can be observed in the combustion duration. The chemical structure of PODE plays an important role in reducing soot formation.

With the rapid development of LED technology, the lighting effect of LED lamp belt in the atmosphere lamp has been unable to meet the needs of the modern market, and the COB lamp belt has gradually entered the public field of vision due to its advantages of high light uniformity, low junction temperature and large luminous Angle. In this paper, the stability, light uniformity, junction temperature and other performance of the two kinds of lamp belt under different color temperature and different power were analyzed by setting contrast experiment. It was found that LED lamp belt had uniform light and poor heat dissipation effect due to chip distribution characteristics, and it would accumulate a lot of heat and appear color shift phenomenon when continuously lit. The COB lamp belt can not only maximize the heat dissipation performance of the substrate, but also achieve more uniform light distribution by the way of multi-chip arrangement. The relationship between the thermal performance and chip spacing is analyzed by optimizing the design, which plays a guiding role in the actual production of the lamp belt

A fixed-time trajectory tracking control of autonomous surface vessels (ASVs) subject to unmeasured speed is studied in this work. By using the homogeneity-based Lyapunov method, the unknown system states, including the unmeasured speed and lumped disturbances, are estimated by using a novel extended state observer (ESO) within fixed time. Subsequently, using the estimated states, the task of fixed-time tracking is completed with the aid of a newly proposed output-constrained power integrator method, which makes the vessel position and heading strictly within the predefined output constraints, and the tracking errors can be reduced to a range of zero under the continuous control action. The practical fixed-time stability (FTS) of the closed-loop system is analyzed in the sense of Lyapunov, while the output constraints can be well maintained during maneuvering. Finally, the ascendancy of the designed scheme is exhibited by simulation comparisons.

Objective. To explore the potential link between gut damage and proinflammatory cytokines in heroin-dependent patients. Methods. We retrospectively analyzed and compared partial blood counts and biomarkers of intestinal injury and their potential correlations in 38 male heroin abuse patients and 29 healthy male participants. In addition, we compared and assessed proinflammatory cytokines and immune cells in 10 heroin abuse patients and 10 healthy participants. Results. Neutrophil counts, platelets/lymphocytes (PLR), neutrophils/lymphocytes (NLR), gut injury biomarkers, and proinflammatory cytokines, CD19+B in patients compared with healthy subjects’ cells increased significantly. The number of lymphocytes, CD3 CD4 T cells, and CD3 CD8 T cells decreased in patients compared to healthy individuals. When distinguishing between heroin addicts and healthy people, ROC/AUC analysis showed that a cutoff of 142.42 for PLR and 2.18 for NLR yielded a sensitivity of 65% and 85% and a specificity of 96.5% and 89.7%, respectively (p=0.001,p<0.001). For predicting intestinal injury, ROC/AUC analysis showed that a cutoff of 135.7 for PLR and 0.15 for NLR yielded a sensitivity of 52% and 60% and a specificity of 82% and 86.4%, respectively (p=0.003,p=0.009). Male heroin addicts are subject to intestinal injury and present with increased proinflammatory cytokine levels. Conclusion. NLR and PLR are possible indirect biomarkers for heroin dependence based on intestinal injury.

Panicle development is crucial to increase the grain yield of rice (Oryza sativa). The molecular mechanisms of the control of panicle development in rice remain unclear. In this study, we identified a mutant with abnormal panicles, termed branch one seed 1-1 (bos1-1). The bos1-1 mutant showed pleiotropic defects in panicle development, such as the abortion of lateral spikelets and the decreased number of primary panicle branches and secondary panicle branches. A combined map-based cloning and MutMap approach was used to clone BOS1 gene. The bos1-1 mutation was located in chromosome 1. A T-to-A mutation in BOS1 was identified, which changed the codon from TAC to AAC, resulting in the amino acid change from tyrosine to asparagine. BOS1 gene encoded a grass-specific basic helix–loop–helix transcription factor, which is a novel allele of the previously cloned LAX PANICLE 1 (LAX1) gene. Spatial and temporal expression profile analyses showed that BOS1 was expressed in young panicles and was induced by phytohormones. BOS1 protein was mainly localized in the nucleus. The expression of panicle development-related genes, such as OsPIN2, OsPIN3, APO1, and FZP, was changed by bos1-1 mutation, suggesting that the genes may be the direct or indirect targets of BOS1 to regulate panicle development. The analysis of BOS1 genomic variation, haplotype, and haplotype network showed that BOS1 gene had several genomic variations and haplotypes. These results laid the foundation for us to further dissect the functions of BOS1.

Reverse water gas shift (RWGS) reaction is a crucial process in CO2 utilization. Herein, Ni- and NiCe-containing hexagonal mesoporous silica (Ni-HMS and NiCe-HMS) catalysts were synthesized using an in-situ one-pot method and applied for RWGS reaction. At certain reaction temperatures 500-750 °C, Ni-HMS samples displayed a higher selectivity to the preferable CO than that of conventionally impregnated Ni/HMS catalyst. This could be originated from the smaller NiO nanoparticles over Ni-HMS catalyst. NiCe-HMS exhibited higher activity compared to Ni-HMS. The catalysts were characterized by means of TEM, XPS, XRD, H2-TPR, CO2-TPD, EPR and N2 adsorption-desortion technology. It was found that introduction of Ce created high concentration of oxygen vacancies, served as the active site for activating CO2. Also, this work analyzed the effect of the H2/CO2 molar ratio on the best NiCe-HMS. When reaction gas H2/CO2 molar ratio was 4 significantly decreased the selectivity to CO at low temperature, but triggered a higher CO2 conversion which is close to the equilibrium.

Image quality assessment (IQA) plays a central role in many image processing algorithms and systems. Although many popular IQA models achieves high performance on existing released databases, they are still not well accepted in practical applications due to the not-always satisfactory accuracy on real-world data and situations. In this paper, we revisit the IQA research, and point out an ignored but interesting problem in IQA: the coarse-grained (i.e., when quality variation is sufficiently big, as the setting of most IQA databases up to date) statistical results evaluated on existing databases mask the fine-grained differentiation. Accordingly, we present a survey on image quality assessment from a new perspective: fine-grained image quality assessment (FG-IQA). Recent FG-IQA research on five major kinds of images is introduced, and some popular IQA methods are analyzed from FG-IQA perspective. The potential problems for current IQA research based on existing coarse-grained databases are analyzed and the necessity of more FG-IQA research is justified. Finally, we discuss some challenges and possible directions for future works in FG-IQA.

Recently, deep learning for video coding, such as deep predictive coding, deep transform coding, and deep in-loop filtering, has been an emerging research area. The coding gain of hybrid coding framework could be extensively promoted by the data-driven models. However, previous deep coding tools especially deep in-loop filtering mainly consider the performance improvement while pay less attention to the reliability, usability, and adaptivity of the networks. In this article, a nested-residual guided convolutional neural network (NR-CNN) structure with cascaded global shortcut and configurable residual blocks is proposed for in-loop filtering. By taking advantage of the correlation between different color components, we further extend the NR-CNN by utilizing luminance as textural and structural guidance for chrominance filtering, which significantly improves the filtering performance. To fully exploit the proposed network into codec integration, we subsequently introduce an efficient and adaptive framework consisting of an adaptive granularity optimization and a parallel inference pipeline for deep learning based filtering. The former contributes to the coding performance improvement through an adaptive decision-making based on rate-distortion analysis at various granularities. The latter reduces the running time of network inference. The extensive experimental results show the superiority of the proposed method, achieving 8.2%, 14.9%, and 13.2% BD-rate savings on average under random access (RA) configuration. Meanwhile, the proposed method also obtains better subjective quality.

Rate control is an important technique for practical video coding applications by allocating suitable quantization parameters (QPs) for CTUs or frames to satisfy bandwidth requirement. However, rate control may result in video quality fluctuation because neighboring regions are quantized into different quality levels, which may affect viewing experience seriously, especially at low bitrate scenarios. In this paper, we propose a novel CTU (Coding Tree Unit) level rate control approach for intra frame of the high efficiency video coding (HEVC) standard with consistent perceptual quality. In particular, a new perceptual hyperbolic rate distortion (R-D) optimization method is proposed for rate control with the constraint of constant perceptual quality, which shows good compatibility with other R-D models. Herein, a novel CTU level perceptual distortion model is presented to achieve the optimal rate control of intra frames for HEVC, where the Lagrange multiplier is adjusted by taking advantage of the different perceptual characteristics to achieve the minimal perceptual distortion variation across CTUs within one frame. Extensive experiments prove that the proposed algorithm outperforms the state-of-the-art CTU level rate control methods by considering R-D performance. In addition, subjective simulation results demonstrate that our algorithm further improves the visual quality of compressed videos by reducing perceptual quality fluctuation.

Triple-negative breast cancer (TNBC) has a high propensity for organ-specific metastasis. However, the underlying mechanisms are not well understood. Here we show that the primary TNBC tumor-derived C-X-C motif chemokines 1/2/8 (CXCL1/2/8) stimulate lung-resident fibroblasts to produce the C-C motif chemokines 2/7 (CCL2/7), which, in turn, activate cholesterol synthesis in lung-colonizing TNBC cells and induce angiogenesis at lung metastatic sites. Inhibiting cholesterol synthesis in lung-colonizing breast tumor cells by pulmonary administration of simvastatin-carrying HER3-targeting nanoparticles reduces angiogenesis and growth of lung metastases in a syngeneic TNBC mouse model. Our findings reveal a novel, chemokine-regulated mechanism for the cholesterol synthesis pathway and a critical role of metastatic site-specific cholesterol synthesis in the pulmonary tropism of TNBC metastasis. The study has implications for the unresolved epidemiological observation that use of cholesterol-lowering drugs has no effect on breast cancer incidence but can unexpectedly reduce breast cancer mortality, suggesting interventions of cholesterol synthesis in lung metastases as an effective treatment to improve survival in individuals with TNBC.

Compactly representing visual information plays a fundamental role in optimizing the ultimate utility of myriad visual data-centered applications. Numerous approaches have been proposed to efficiently compress the texture and visual features for human visual perception and machine intelligence, respectively; however, much less work has been dedicated to studying the interactions between them. Here, we investigate the integration of feature and texture compression and show that a universal and collaborative visual information representation can be achieved in a hierarchical way. In particular, we study feature and texture compression in a scalable coding framework, where the base layer serves as the deep learning feature and the enhancement layer targets to perfectly reconstruct the texture. Based on the strong generative capability of deep neural networks, the gap between the base feature layer and enhancement layer is further filled with feature-level texture reconstruction, with the goal of further constructing texture representations from features. As such, the residuals between the original and reconstructed texture could be further conveyed in the enhancement layer. To improve the efficiency of the proposed framework, the base layer neural network is trained in a multitask manner such that the learned features enjoy both high-quality reconstruction and high-accuracy analysis. The framework and optimization strategies are further applied in face image compression, and promising coding performance has been achieved in terms of both rate-fidelity and rate-accuracy evaluations.

In-loop filters have attracted increasing attention due to the remarkable noise-reduction capability in the hybrid video coding framework. However, the existing in-loop filters in Versatile Video Coding (VVC) mainly take advantage of the image local similarity. Although some non-local based in-loop filters can make up for this shortcoming, the widely-used unsupervised parameter estimation method by non-local filters limits the performance. In view of this, we propose a deformable Wiener Filter (DWF). It combines the local and non-local characteristics and supervisedly trains the filter coefficients based on the Wiener Filter theory. In the filtering process, local adjacent samples and non-local similar samples are first derived for each sample of interest. Then the to-be-filtered samples are classified into specific groups based on the patch-level noise and sample-level characteristics. Samples in each group share the same filter coefficients. After that, the local and non-local reference samples are adaptively fused based on the classification results. Finally, the filtering operation with outlier data constraints is conducted for each to-be-filtered sample. Moreover, the performance of the proposed DWF is analyzed with different reference sample derivation schemes in detail. Simulation results show that the proposed approach achieves 1.16%, 1.92%, and 2.67% bit-rate savings on average compared to the VTM-11.0 for All Intra, Random Access, and Low-Delay B configurations, respectively.

In the latest video coding standard, namely Versatile Video Coding (VVC), more directional intra modes and reference lines have been utilized to improve prediction efficiency. However, complex content still cannot be predicted well with only the adjacent reference samples. Although nonlocal prediction has been proposed to further improve the prediction efficiency in existing algorithms, explicit signalling or matching error potentially limits the coding efficiency. To address these issues, we propose a joint local and nonlocal progressive prediction scheme, targeting at improving nonlocal prediction accuracy without additional signalling. Specifically, template matching based prediction (TMP) is conducted firstly to derive an initial nonlocal predictor. Based on the first prediction and previously decoded reconstruction information, a local template, including inner textures and neighboring reconstruction, is carefully designed. With the local template involved in nonlocal matching process, a more accurate nonlocal predictor can be found progressively in the second prediction. Finally, the coefficients from the two predictions are fused and transmitted in bitstreams. In this way, more accurate nonlocal predictor can be derived implicitly with local information instead of being explicitly signalled. Experimental results on the reference software VTM-9.0 of VVC show that the method achieves 1.02% BD-Rate reduction for natural sequences and 2.31% BD-Rate reduction for screen content videos under all intra (AI) configuration.

Geometric partitioning has attracted increasing attention by its remarkable motion field description capability in the hybrid video coding framework. However, the existing geometric partitioning (GEO) scheme in Versatile Video Coding (VVC) causes a non-negligible burden for signaling the side information. Consequently, the coding efficiency is limited. In view of this, we propose a spatio-temporal correlation guided geometric partitioning (STGEO) scheme to efficiently describe the object information in the motion field of video coding. The proposed method can economize the bits consumed for side information signaling, including the partitioning mode and motion information. We firstly analyze the characteristics of partitioning mode decision and motion vector selection in a statistically-sound way. Based on the observed spatio-temporal correlation, we design a mode prediction and coding method to reduce the overhead for representing the above mentioned side information. The main idea is to predict the STGEO modes and motion candidates that have higher selection possibilities, which can guide the entropy coding, i.e., representing the predicted high-probability modes and motion candidates with fewer bits. In particular, the high-probability STGEO modes are predicted based on the edge information and history modes of adjacent STGEO-coded blocks. The corresponding motion information is represented by the index in a merge candidate list, which is adaptively inferred based on the off-line trained merge candidate selection probability. Simulation results show that the proposed approach achieves 0.95% and 1.98% bit-rate savings on average compared to VTM-8.0 without GEO for Random Access and Low-Delay B configurations, respectively.

We reported the exploration of reliable and flexible supercapacitors toward wide-temperature operation based on SiC/CNT composite films, which have long-term cycle stability up to 14 000 cycles under variated temperatures between −30 and 50 °C.

Background: Collecting duct carcinoma (CDC) is a rare renal tumor, originating from the distal collecting duct. CDC rarely presents as a primary tumor outside the renal system. Case presentation: In this study, we report a rare case of collecting duct carcinoma, with an initial presentation of retroperitoneal lymph node metastasis, and no identifiable primary renal tumor on CT, at the time of diagnosis. The patient was a 64-year-old man presenting with lower back pain. Preoperative CT showed a round, soft tissue mass, measuring 6.7×4.4×3.3 cm, in the left retroperitoneum with no exact occupying lesion in the left kidney. Clinically, ectopic pheochromocytoma was considered to be a differential diagnosis, and tumor resection was performed. Postoperative pathological results demonstrated that the mass was a fused lymph node, and the tumor cells were destroying the structure. The final diagnosis was lymph node metastatic collecting duct carcinoma, by histology and immunohistochemistry. No further treatment was performed as no space occupying lesion was found in the kidney. Three months later, CT was reexamined, and a mass of 3.6 cm in diameter, was found in the lower left kidney, along with multiple soft tissue masses, in the left renal hilum. Considering recurrence or metastasis, the patient was recommended to undergo surgical treatment, but the patient refused. Four months later, CT was re-examined. The tumor had rapidly progressed but the patient refused treatment again. As per the author's press release (eleven months after the first discovery), the patient is still alive. Conclusion: CDC is a rare malignant renal carcinoma, with a high chance of rapid progress, regional lymph nodes involvement and metastasis. It presents diagnostic challenges to clinicians and pathologists, particularly, in the absence of radiographically detectable intrarenal lesions. Definite diagnosis is based on pathological examination combined with immunohistochemical staining.

In this paper, the hybrid-scale bifurcation evolutionary phenomena in one-cycle controlled (OCC) single-inductor dual-output (SIDO) Buck DC–DC converter are investigated thoroughly. First, the single-cycle and crossing-cycle iterative mapping models of the OCC SIDO converter in four operation modes are derived to describe the dynamical information under different topology switching sequences. Second, the evolutionary trend of hybrid-scale bifurcation behaviors is identified by numerical simulation. Third, bifurcation analysis based on the proposed models is performed by the trajectories of the eigenvalues, and significantly, the occurrence mechanism of hybrid-scale bifurcations in the evolutionary phenomena is explained in detail. And then, the participation factor is used to analyze the relationship between the variables and the eigenvalues. Especially, the relative parameter sensitivity of the iterative mapping model is proposed to evaluate the influence of such key parameters as load resistors on the eigenvalues so that plenty of design-oriented information will be provided to avoid the hybrid-scale bifurcation behaviors. Finally, PSpice circuit simulation and hardware experiments are carried out to verify the analysis results.

Aggregation-induced emission (AIE) molecules that can avoid the aggregation-caused quenching (ACQ) effect and break the concentration limit have been widely used for biosensing. Similar to fluorescence dyes, AIE molecules can be chemiexcited simply by a peroxyoxalate-based chemiluminescence (CL) reaction, but the hydrolysis of peroxyoxalate is often a problem in an aqueous solution. Herein, we report an AIE effect within peroxyoxalate-loaded silica nanoparticles (PMSNs) for an efficient harvest of CL energy as well as alleviation of bis(2,4,5-trichloro-6-carbopentoxyphenyl) oxalate (CPPO) hydrolysis. Peroxyoxalate (i.e., CPPO) and AIE molecules (i.e., 1,2-benzothiazol-2-triphenylamino acrylonitrile, BTPA) were loaded together within the mesoporous silica nanoparticles (MSNs) to synthesize the BTPA-PMSN nanocomposite. The BTPA-PMSNs not only allowed CPPO to be dispersed well in an aqueous solution but also avoided the hydrolysis of CPPO. Meanwhile, the proximity between BTPA and CPPO molecules in the mesopores of MSNs facilitated the BTPA aggregate to harvest the energy from CL intermediates. Hence, the CL system of BTPA-PMSNs can work efficiently in aqueous solutions at a physiological pH. The CL quantum yield of the BTPA-PMSN system was measured to be 9.91 × 10

Chemiluminescence (CL) has been widely used for bioanalysis owing to its high sensitivity, low background and simplicity. However, most of the CL systems need acidic/alkaline conditions or organic solvent to enhance their luminescent efficiency, and the non-physiological conditions can usually lead to the misfunction of biomolecules during biosensing. Herein, we report a highly luminous CL system under physiological conditions based on carbon dots-bis(2-carbopentyloxy-3,5,6-trichlorophenyl) oxalate (CDs-CPPO) micelles, and further used it in biosensing application. In the CL system, the amphiphilic surfactant packed CPPO and hydrophobic CDs together to form CDs-CPPO micelles. Such micelles solution not only isolated the CPPO from water to prevent its hydrolysis but also made the close proximity between CPPO and CDs, thus significantly enhancing the CDs quantum yield. The CL quantum yield was calculated to be 5.26 × 10−4 einsteins/mol, about 200-fold higher than that of the most commonly used luminol CL system. The oxidases (e.g., glucose oxidase) were tested to be susceptible to the organic solvent and non-physiological pH. Hence, the CL system was used for the detection of oxidase substrates (exemplified by glucose) in serum samples, and the limit of detection was as low as 8.4 nmol/L. The highly luminous CL system that can work under physiological conditions is promising for biosensing applications

The discovery of new edible flowers that are nontoxic, innocuous flowers having human health benefits, surveys of their phytochemicals and utilization are of great scientific and commercial interest. Dendrobium officinale and Dendrobium devonianum are precious Traditional Chinese Medicine. During the massive commercial cultivation, a lot of flowers were produced and certified as edible flowers, and the phytochemical profiles and bioactivities warrant evaluate. The present study aimed to investigate the phytochemicals and antioxidative activities in flowers of D. officinale (DOF) and D. devonianum (DDF). In total, 474 metabolites were identified using a widely targeted metabonomics method, 16 amino acids and 6 flavonoids were measured using high‐performance liquid chromatography (HPLC), and 8 fatty acids were detected using gas chromatography–mass spectrometry (GC‐MS). Both flowers contained various amino acids, including 7 essential amino acids, diverse flavonoids, especially quercetin, kaempferol and their derivatives, and high levels of methyl linoleate and methyl linolenate. The relative levels of quercetin, kaempferol and their glycosides were higher in DDF than in DOF, whereas the relative levels of several flavonoids C‐glycosides were high in DOF. Ethanol extracts of both DOF and DDF showed antioxidative capacities including the scavenging of 1,1‐diphenyl‐2‐picrylhydrazyl and hydroxyl radicals. Both edible flowers contained flavonoids, amino acids, and fatty acids and have antioxidative activities, which should be explored for use in functional foods and pharmaceuticals.

Contrast plays an important role in human perception of image quality. In this paper, we propose a metric for no-reference quality assessment of contrast-changed images by using a novel semi-supervised robust PCA , which can realize feature selection and denoising simultaneously, guided by the available supervisory information. To select features adaptively, the information-oriented features (e.g. entropy and natural scene statistics) and appearance-oriented features (e.g. colorfulness) are adopted. The proposed model is formulated as a constraint optimization problem , which is further casted to a convex problem and solved via augmented Lagrangian multiplier method. Extensive experimental results on CCID2014, CSIQ, SIQAD and TID2013 databases show that the proposed semi-supervised image quality metric based on robust PCA (SIQMR) provides a more accurate prediction than other metrics on the human perception of contrast variations.

Background: Tyrosine kinase inhibitors (TKI) in combination with programmed cell death-1 (PD-1) inhibitors become the potential treatment modality for patients undergoing unresectable hepatocellular carcinoma (uHCC) in the first-line setting. However, the efficacy and safety of this combination regimen in patients after sorafenib failure remains unclear.Methods: Participants in this study included patients with uHCC after sorafenib failure who received TKI monotherapy (TKI group) or TKI combined with PD-1 inhibitors therapy (combination group) in our center from July 2018 to July 2021. The overall survival (OS) was used to be the primary efficacy endpoint, while progression-free survival (PFS), objective response rate (ORR), and disease control rate (DCR) were applied to be secondary endpoints. In addition, the adverse events are recorded and evaluated.Results: Among the 92 patients contained in this work, 50 patients were categorized into the TKI group, while 42 patients were in the combination group. There existed no evident differences between the two groups concerning the ORR (8.0% vs. 9.5%, p = 1.000). However, the DCR in the combined group was better in relative to that in the TKI group (71.4% vs. 50.0%, p = 0.037). In comparison with the TKI group, it was found that the combination group presented notably better median PFS (8.1 months vs. 4.7 months, p = 0.005) and median OS (21.9 months vs. 16.6 months, p = 0.042). According to multivariate analysis, PFS (HR 0.5, 95% CI: 0.3–0.8, p = 0.005) and OS (HR 0.5, 95% CI: 0.3–1.0, p = 0.051) were improved in the combination group in relative to the TKI group after the adjustment for some risk factors. Additionally, the incidence rates of grade ≥1 adverse event in the TKI group and the combination group were 96.0% and 97.6%, respectively. The most normal adverse event in the TKI group was neutropenia (n = 24,48.0%) and the combination group was hypoalbuminemia (n = 23,54.8%). All of these adverse events improved after symptomatic treatment, and no new toxic events were found to occur.Conclusion: TKI combined with PD-1 inhibitors showed better prognosis with manageable toxicity in uHCC patients after sorafenib failure compared with TKI monotherapy.