U2-Net and ResNet50-Based Automatic Pipeline for Bacterial Colony Counting.

In this paper, an automatic colony counting system based on an improved image preprocessing algorithm and convolutional neural network (CNN)-assisted automatic counting method was developed. Firstly, we assembled an LED backlighting illumination platform as an image capturing system to obtain photographs of laboratory cultures. Consequently, a dataset was introduced consisting of 390 photos of agar plate cultures, which included 8 microorganisms. Secondly, we implemented a new algorithm for image preprocessing based on light intensity correction, which facilitated clearer differentiation between colony and media areas. Thirdly, a U2-Net was used to predict the probability distribution of the edge of the Petri dish in images to locate region of interest (ROI), and then threshold segmentation was applied to separate it. This U2-Net achieved an F1 score of 99.5% and a mean absolute error (MAE) of 0.0033 on the validation set. Then, another U2-Net was used to separate the colony region within the ROI. This U2-Net achieved an F1 score of 96.5% and an MAE of 0.005 on the validation set. After that, the colony area was segmented into multiple components containing single or adhesive colonies. Finally, the colony components (CC) were innovatively rotated and the image crops were resized as the input (with 14,921 image crops in the training set and 4281 image crops in the validation set) for the ResNet50 network to automatically count the number of colonies. Our method achieved an overall recovery of 97.82% for colony counting and exhibited excellent performance in adhesion classification. To the best of our knowledge, the proposed "light intensity correction-based image preprocessing→U2-Net segmentation for Petri dish edge→U2-Net segmentation for colony region→ResNet50-based counting" scheme represents a new attempt and demonstrates a high degree of automation and accuracy in recognizing and counting single-colony and multi-colony targets.

Research on system of ultra-flat carrying robot based on improved PSO algorithm.

Ultra-flat carrying robots (UCR) are used to carry soft targets for functional safety road tests of intelligent driving vehicles and should have superior control performance. For the sake of analyzing and upgrading the motion control performance of the ultra-flat carrying robot, this paper develops the mathematical model of its motion control system on the basis of the test data and the system identification method. Aiming at ameliorating the defects of the standard particle swarm optimization (PSO) algorithm, namely, low accuracy, being susceptible to being caught in a local optimum, and slow convergence when dealing with the parameter identification problems of complex systems, this paper proposes a refined PSO algorithm with inertia weight cosine adjustment and introduction of natural selection principle (IWCNS-PSO), and verifies the superiority of the algorithm by test functions. Based on the IWCNS-PSO algorithm, the identification of transfer functions in the motion control system of the ultra-flat carrying robot was completed. In comparison with the identification results of the standard PSO and linear decreasing inertia weight (LDIW)-PSO algorithms, it indicated that the IWCNS-PSO has the optimal performance, with the number of iterations it takes to reach convergence being only 95 and the fitness value being only 0.117. The interactive simulation model was constructed in MATLAB/Simulink, and the critical proportioning method and the IWCNS-PSO algorithm were employed respectively to complete the tuning and optimization of the Proportional-Integral (PI) controller parameters. The results of simulation indicated that the PI parameters optimized by the IWCNS-PSO algorithm reduce the adjustment time to 7.99 s and the overshoot to 13.41% of the system, and the system is significantly improved with regard to the control performance, which basically meets the performance requirements of speed, stability, and accuracy for the control system. In conclusion, the IWCNS-PSO algorithm presented in this paper represents an efficient system identification method, as well as a system optimization method.

Estrogen receptor α-mediated signaling inhibits type I interferon response to promote breast cancer.

Estrogen receptor α (ERα) is an important driver and therapeutic target in approximately 70% of breast cancers. How ERα drives breast carcinogenesis is not fully understood. In this study, we show that ERα is a negative regulator of type I interferon (IFN) response, which is critical for breast carcinogenesis. Activation of ERα by its natural ligand estradiol inhibits IFN-ß-induced transcription of downstream IFN-stimulated genes (ISGs), whereas deficiency of ERα or stimulation with its antagonist fulvestrant has opposite effects. Mechanistically, ERα inhibits type I IFN response by two distinct mechanisms. ERα induces expression of the histone 2A variant H2A.Z, which restricts engagement of the IFN-stimulated gene factor 3 (ISGF3) complex at the ISG promoters. ERα also interacts with STAT2, which leads to disruption of the ISGF3 complex. These two events mutually lead to transcriptional inhibition of ISGs induced by type I IFNs. In a xenograft mouse tumor model, fulvestrant enhances the ability of IFN-ß to suppress ERα+ breast tumor growth. Consistently, clinical data suggests that ERα+ breast cancer patients with higher levels of ISGs exhibit an increased survival rate. Our findings suggest that ERα inhibits type I IFN response via two distinct mechanisms to promote breast cancer.

Endocytosis triggers V-ATPase-SYK-mediated priming of cGAS activation and innate immune response.

The current view of nucleic acid-mediated innate immunity is that binding of intracellular sensors to nucleic acids is sufficient for their activation. Here, we report that endocytosis of virus or foreign DNA initiates a priming signal for the DNA sensor cyclic GMP-AMP synthase (cGAS)-mediated innate immune response. Mechanistically, viral infection or foreign DNA transfection triggers recruitment of the spleen tyrosine kinase (SYK) and cGAS to the endosomal vacuolar H+ pump (V-ATPase), where SYK is activated and then phosphorylates human cGASY214/215 (mouse cGasY200/201) to prime its activation. Upon binding to DNA, the primed cGAS initiates robust cGAMP production and mediator of IRF3 activation/stimulator of interferon genes-dependent innate immune response. Consistently, blocking the V-ATPase-SYK axis impairs DNA virus- and transfected DNA-induced cGAMP production and expression of antiviral genes. Our findings reveal that V-ATPase-SYK-mediated tyrosine phosphorylation of cGAS following endocytosis of virus or other cargos serves as a priming signal for cGAS activation and innate immune response.

Integrated strategy for widely targeted metabolome characterization of Peucedani Radix.

Herbal medicines (HMs) are widely recognized as extremely complicated matrices, resulting in a great challenge for the existing analytical approaches to characterize the widely targeted metabolome. The primary obstacles include high-level structural diversity, broad concentration range, large polarity span, insufficient authentic compounds and frequent occurrences of isomers, even enantiomers. Here, we aimed to propose an integrated strategy being able to circumvent the technical barriers, and a well-known HM namely Peucedani Radix was employed to illustrate and justify the applicability. Regarding qualitative analysis, the hydrophilic metabolites were detected with HILIC-predictive multiple-reaction monitoring mode, and structurally identified by matching predefined identities with authentic compounds or information archived in relevant databases. After RPLC-MS/MS measurement, full collision energy ramp-MS2 spectrum in combination with quantum structural calculation was applied to confirmatively identify those less polar components, mainly angular-type pyranocoumarins (APs). For quantitative analysis, achiral-chiral RPLC/HILIC was configured for chromatographic separations because the analytes spanned a large polarity range and involved many enantiomers. A quasi-content concept was employed for comprehensively relative quantitation through constructing a so-called universal metabolome standard (UMS) sample and building calibration curves by assaying serial diluted UMS solutions. Consequently, high-confidence structural annotation and relatively quantitative analysis were achieved for 103 compounds, in total. After multivariate statistical analysis, some APs, e.g., (3'S)-praeruptorin A, (3'S)-praeruptorin B, (3'S)-praeruptorin E, as well as several primary metabolites were screened out as the prominent contributors for inter-batch variations. Together, current study shows a promising strategy enabling widely targeted metabolomics of, but not limited to, HMs.

Modulation of innate immune response to viruses including SARS-CoV-2 by progesterone.

Whether and how innate antiviral response is regulated by humoral metabolism remains enigmatic. We show that viral infection induces progesterone via the hypothalamic-pituitary-adrenal axis in mice. Progesterone induces downstream antiviral genes and promotes innate antiviral response in cells and mice, whereas knockout of the progesterone receptor PGR has opposite effects. Mechanistically, stimulation of PGR by progesterone activates the tyrosine kinase SRC, which phosphorylates the transcriptional factor IRF3 at Y107, leading to its activation and induction of antiviral genes. SARS-CoV-2-infected patients have increased progesterone levels, and which are co-related with decreased severity of COVID-19. Our findings reveal how progesterone modulates host innate antiviral response, and point to progesterone as a potential immunomodulatory reagent for infectious and inflammatory diseases.

[Feasibility Study of Ultra-High-Resolution Low-Dose Temporal Bone CT with 1 024×1 024 Reconstruction Matrix Size].

OBJECTIVE: To investigate the feasibility of low-dose CT scan of the temporal bone combined with reconstruction matrix size of 1 024×1 024 and the effect of the reconstruction matrix size on image quality. METHODS: Normal-dose and low-dose bilateral temporal bone CT scans were performed on twelve adult male cadaveric skull specimens using the 160-slice multi-detector CT scanning of United Imaging Healthcare. Normal-dose CT images were reconstructed with matrix sizes of 512×512 and 1 024×1 024, while low-dose CT images were reconstructed with the matrix size of 1 024×1 024. CT value, noise, signal-to-noise ratio, contrast-to-noise ratio, the visualization scoring of 15 anatomical structures of the temporal bone, and the result of three-dimensional reconstruction of the ossicular chain were compared among the three groups. RESULTS: The radiation dose of low-dose CT scanning was reduced by about 50% compared with that of normal-dose CT. There was no significant difference in CT values of air, soft tissues and bones among the three groups. Low-dose temporal bone CT with the matrix size of 1 024×1 024 had higher noise, but much better visualization of temporal bone structure than the normal-dose temporal bone CT with matrix size of 512×512. Both the three-dimensional reconstructions of normal-dose and low-dose 1 024×1 024 matrix images were satisfactory and showed no significant difference. The morphology, size and relative position of malleus, incus, stapes, cochlea, and labyrinth, as well as the location of the ossicular chain in the cranium were all clearly displayed. CONCLUSION: Low-dose temporal bone CT with the matrix size of 1 024×1 024 can be used to effectively reduce the radiation dose and significantly improve the spatial resolution and the visualization of the temporal bone anatomical structures compared with the normal-dose temporal bone CT with a matrix size of 512×512.

[Rapid chemome profiling of Cistanche salsa using DI-MS/MS~(ALL)].

The current study aims to rapidly and comprehensively profile the chemical composition of Cistanche salsa using direct infusion coupled with MS/MS~(ALL)(DI-MS/MS~(ALL)). The C. salsa extract was directly imported into electrospray ionization(ESI) source of quadrupole time-of-flight(Q-TOF) mass spectrometer with an infusion pump at a flow rate of 10 µL·min~(-1). Acquisition program was applied under negative ionization polarity to collect one MS~1 spectrum(m/z 50-1 200), followed by 1 150 MS~2 spectra with precursor isolation window(m/z 1) amongst mass range m/z 50-1 200. After each MS~2 spectrum was matched to its precursor ion, putative identification was conducted through matching mass spectral data with literature and database. A total of 31 components were identified from C. salsa, including 9 phenylethanoid glycosides, 2 iridoids, 4 saccharides, 9 organic acids, and 7 other compounds, similar to those from C. tubulosa and C. deserticola. In conclusion, DI-MS/MS~(ALL), a facile and reliable analytical tool, can be employed for qualitative analysis of chemical constituents in C. salsa. The research offers a promising strategy to achieve rapid chemome profiling of herbal medicine and provides an alternative source of Cistanches Herba.

[Rapid chemome profiling of chemical components of Lonicerae Japonicae Flos using DI-MS/MS~(ALL)].

A new method of MS/MS~(ALL) was designed to sequentially record a MS~2 spectrum at each unit mass window through gas phase fractionation concept, so as to offer an opportunity for universal MS~2 spectral recording with direct infusion(DI). As a proof-of-concept, DI-MS/MS~(ALL) was applied for rapid chemome profiling of a famous herbal medicine named Lonicerae Japonicae Flos. After each MS~2 spectrum was correlated to its precursor ion, the structural annotation was conducted by applying well-defined mass cracking rules, matching the mass spectral data with literatures and referring to those accessible databases. As a result, a total of 54 components were identified from Lonicerae Japonicae Flos extract, including 21 phenolic acids, 13 flavonoids, 12 iridoids, 4 triterpenoids and 4 other compounds. Therefore, DI-MS/MS~(ALL) is a powerful tool for comprehensive, rapid qualitative analysis of chemical profiles of traditional Chinese medicine and other chemical components of complex systems.

VRK2 is involved in the innate antiviral response by promoting mitostress-induced mtDNA release.

Mitochondrial stress (mitostress) triggered by viral infection or mitochondrial dysfunction causes the release of mitochondrial DNA (mtDNA) into the cytosol and activates the cGAS-mediated innate immune response. The regulation of mtDNA release upon mitostress remains uncharacterized. Here, we identified mitochondria-associated vaccinia virus-related kinase 2 (VRK2) as a key regulator of this process. VRK2 deficiency inhibited the induction of antiviral genes and caused earlier and higher mortality in mice after viral infection. Upon viral infection, VRK2 associated with voltage-dependent anion channel 1 (VDAC1) and promoted VDAC1 oligomerization and mtDNA release, leading to the cGAS-mediated innate immune response. VRK2 was also required for mtDNA release and cGAS-mediated innate immunity triggered by nonviral factors that cause Ca2+ overload but was not required for the cytosolic nucleic acid-triggered innate immune response. Thus, VRK2 plays a crucial role in the mtDNA-triggered innate immune response and may be a potential therapeutic target for infectious and autoimmune diseases associated with mtDNA release.

Simultaneous determination of eight tryptic peptides in musk using high-performance liquid chromatography coupled with tandem mass spectrometry.

In comparison of herbal medicines, less attention has been paid onto animal medicines, partially attributing to the protein-enriched property. Particularly, it is still challenging to conduct quality evaluation for the animal medicines because of the lack of a fit-for-purpose analytical tool. Herein, an attempt was made to propose a workflow allowing the quality assessment of animal medicines by LC-MS/MS, and musk that is one of the most precious traditional Chinese medicines was employed as a representative case for utility illustration. After the extraction of protein from musk with a well-defined protocol, tryptic digestion was conducted to hydrolyze proteins into peptides, and the peptide-enriched sample was subjected to nanoLC-Orbitrap MS measurement. The tandem mass spectral dataset was retrieved in Human Swiss-Prot FASTA database, and the sequences together with the sources of 733 tryptic peptides, in total, were annotated. Because of the abundant distributions, eight peptides were chosen as the analytes for quantitative measurements, and their quantitative MS parameters, such as ion transitions and collision energies, were rapid optimized in an authentic compound-free manner using online energy-resolved MS (ER-MS). On the other side, the annotated peptides were structurally consolidated via synthesizing reference peptides. When the synthetic peptides were applied for parameter optimization with the authentic compound-dependent manner, the values were almost identical with those from online ER-MS measurements. After being validated with diverse assays, the developed method was applied for the simultaneous determination of eight peptides in 28 batches of musk samples, including captive (C1-C18) and wild ones (W1-W10). Significant differences took place for the content patterns of concerned tryptic peptides between the captive and wild musk samples. Trace distributions occurred for DVDAAYMNK in most batches. Captive samples were rich of QSLEASLAETEGR, TLLDIDNTR, and EVATNSELVQSGK, whereas wild samples were able to accumulate YLGYLEQLLR. Together, the present study provided a meaningful approach for the quality evaluation of musk, as well as other peptide-enriched animal medicines, even if the absences of authentic peptides.

Integrated Strategy Drives Direct Infusion-Tandem Mass Spectrometry as an Eligible Tool for Shotgun Pseudo-Targeted Metabolomics of Medicinal Plants.

Direct infusion (DI) has an extraordinary high-throughput advantage. Pseudo-targeted metabolomics (PTM) has been demonstrated integrating the merits of both nontargeted and targeted metabolomics. Herein, we attempted to implant DI into the PTM concept to configure a new strategy allowing shotgun PTM. First, a versatile MS/MSALL program was applied to acquire MS1 and MS2 spectra. Second, online energy-resolved MS (online ER-MS) was conducted to obtain breakdown graph as well as optimal collision energy (OCE) for each ion transition paired by precursor ion and the dominant product ion. Third, selected reaction monitoring (SRM) was responsible to output a quantitative dataset with a constant length. Moreover, breakdown graph also served as orthogonal structural evidence when matching MS2 spectra between DI-MS/MS and an in-house library to strengthen structural annotation confidence. To evaluate and illustrate the utility of the new strategy toward shotgun PTM of medicinal plants, in-depth chemome comparison was conducted within three Cistanche species, all of which are edible medicinal plants and playing essential roles for turning the deserts into the oases. A total of 185 variables participated in the quantitative measurement program. Each diagnostic ion pair was featured with an OCE. Significant species differences occurred, and echinacoside, acteoside, isoacteoside, 2'-acetyl-acteoside, tubuloside B, mannitol, sucrose, betaine, malate, as well as choline were found to be confirmative chemical markers offering primary contributions toward the species discrimination. After cross-validation with LC-MS/MS, DI-MS/MS fortified with the new strategy is an eligible tool for shotgun PTM, beyond Cistanche plants.

Is the 0.2%-Strain-Offset Approach Appropriate for Calculating the Yield Stress of Cortical Bone?

The 0.2% strain offset approach is mostly used to calculate the yield stress and serves as an efficient method for cross-lab comparisons of measured material properties. However, it is difficult to accurately determine the yield of the bone. Especially when computational models require accurate material parameters, clarification of the yield point is needed. We tested 24 cortical specimens harvested from six bovine femora in three-point bending mode, and 11 bovine femoral cortical specimens in the tensile mode. The Young's modulus and yield stress for each specimen derived from the specimen-specific finite element (FE) optimization method was regarded as the most ideal constitutive parameter. Then, the strain offset optimization method was used to find the strain offset closest to the ideal yield stress for the 24 specimens. The results showed that the 0 strain offsets underestimated (- 25%) the yield stress in bending and tensile tests, while the 0.2% strain offsets overestimated the yield stress (+ 65%) in three-point bending tests. Instead, the yield stress determined by 0.007 and 0.05% strain offset for bending and tensile loading respectively, can effectively characterize the biomechanical responses of the bone, thereby helping to build an accurate FE model.

Liquid chromatography-three-dimensional mass spectrometry enables confirmative structural annotation of cistanoside F metabolites in rat.

Clarification the existence forms, including prototype and metabolite(s) is the prerequisite for understanding in-depth the therapeutic mechanisms of a given agent, particularly when oral administration. However, it is still a long distance for unambiguous structural identification of metabolites even employing the cutting-edge MS/MS technique, and the determinant obstacle is produced by its inherent isomer-blind disadvantage. To tackle with this drawback, online energy-resolved mass spectrometry (online ER-MS) was introduced to enable isomeric discrimination after that high-resolution MS/MS provided empirical molecular formula as well as substructures. In-depth metabolic characterization of cistanoside F (CF), an effective natural product, was conducted as a proof-of-concept for the new strategy namely three-dimensional MS that was configured by MS1, MS2 and online ER-MS as 1st, 2nd, and 3rd dimensions, respectively. Sensitive metabolite detection was assisted by predictive multiple-reaction monitoring function on Qtrap-MS, and the empirical formulas of all metabolites were calculated from the quasi-molecular ions yielded from IT-TOF-MS. Subsequently, substructures of each metabolite were constructed by combining the calculated element compositions and the well-defined mass fragmentation pathways. Finally, online ER-MS was responsible to generate optimal collision energies for bonds-of-interest, and enabled rational selection among candidate structures. A total of thirteen metabolites were detected and confirmatively identified in rat after oral treatment of CF using LC-3D MS. Acyl-migration, hydrolysis and sulfation played key roles for the metabolic fate of CF. More importantly, LC-3D MS is an eligible tool to achieve confidence-enhanced structural annotation of metabolites in biological matrices because of the unique isomeric differentiation ability from online ER-MS.

Development of Test Equipment for Pedestrian-Automatic Emergency Braking Based on C-NCAP (2018).

In order to evaluate the effectiveness of a pedestrian-automatic emergency braking (PAEB) system on pedestrian protection, a set of PAEB test equipment was developed according to the test requirement of China-New Car Assessment Program (C-NCAP) (2018) in this study. In the aspect of system control strategy, global positioning system (GPS) differential positioning was used to achieve the required measurement and positioning accuracy, the collaborative control between the PAEB test equipment and automated driving robot (ADR) was achieved by wireless communication, and the motion state of the dummy target in the PAEB system was controlled by using the S-shaped-curve velocity control method. Part of the simulations and field tests were conducted according to the scenario requirements specified in C-NCAP (2018). The experimental and simulated results showed that the test equipment demonstrated high accuracy and precision in the process of testing, the dummy target movement was smooth and stable, complying with the requirements of PAEB tests set forth in C-NCAP (2018), and yielding satisfactory results as designed. Subsequently, the performance of the AEB of a vehicle under test (VUT) was conducted and the score for star-rating to evaluate the performance level of AEB calculated. Results indicated the developed test equipment in this study could be used to evaluate the performance of the PAEB system with effectiveness.

Confirmative Structural Annotation for Metabolites of (R)-7,3'-Dihydroxy-4'-methoxy-8-methylflavane, A Natural Sweet Taste Modulator, by Liquid Chromatography-Three-Dimensional Mass Spectrometry.

Flavonoids occupy the largest family of natural products and possess a broad spectrum of health benefits. Their metabolites are sometimes the truly effective molecules in vivo. It is still challenging, however, to unambiguously identify flavonoid metabolites using conventional LC-MS/MS. Herein, we aimed to pursue auxiliary structural clues to m/z values in both MS1 and MS2 spectra through LC coupled to three-dimensional MS (LC-3D MS). MS1, as the first dimension, was in charge of suggesting theoretical molecular formulas, MS2, the as second dimension, was responsible for offering substructures, and exactly, online energy-resolved MS (ER-MS), as the third dimension, provided optimal collision energies (OCEs) that reflected the linkage manners among the substructures. Metabolic characterization of a natural sweet taste modulator, namely, (R)-7,3'-dihydroxy-4'-methoxy-8-methylflavane (DHMMF), was conducted as a proof-of-concept. Extensive efforts, such as full MS1 and MS2 scans on IT-TOF-MS and predictive selected-reaction monitoring mode on Qtrap-MS, were made for in-depth metabolite mining. Seventeen metabolites (M1-M17) were captured from DHMMF-treated biological samples, including 17 (M1-M17), 10 (M4-M9, M11, M13, M14, and M16), and 2 (M5 and M10) metabolites from urine, plasma, and feces, respectively. Their structures were configured by integrating MS1, MS2, and OCE information. Except M10, all metabolites were new compounds. LC-MS/MS-guided chromatographic purification yielded three glucuronyl-conjugated metabolites (M5, M8, and M11), and NMR spectroscopic assays consolidated the structures transmitted from LC-3D MS. Demethylation, glucuronidation, and sulfation occurred as the primary metabolic pathways of DHMMF. Above all, LC-3D MS bridged LC-MS/MS from putatively structural annotation toward confidence-enhanced identification, beyond the metabolite characterization of flavonoids.

UnetDVH-Linear: Linear Feature Segmentation by Dilated Convolution with Vertical and Horizontal Kernels.

Linear feature extraction is crucial for special objects in semantic segmentation networks, such as slot marking and lanes. The objects with linear characteristics have global contextual information dependency. It is very difficult to capture the complete information of these objects in semantic segmentation tasks. To improve the linear feature extraction ability of the semantic segmentation network, we propose introducing the dilated convolution with vertical and horizontal kernels (DVH) into the task of feature extraction in semantic segmentation networks. Meanwhile, we figure out the outcome if we put the different vertical and horizontal kernels on different places in the semantic segmentation networks. Our networks are trained on the basis of the SS dataset, the TuSimple lane dataset and the Massachusetts Roads dataset. These datasets consist of slot marking, lanes, and road images. The research results show that our method improves the accuracy of the slot marking segmentation of the SS dataset by 2%. Compared with other state-of-the-art methods, our UnetDVH-Linear (v1) obtains better accuracy on the TuSimple Benchmark Lane Detection Challenge with a value of 97.53%. To prove the generalization of our models, road segmentation experiments were performed on aerial images. Without data argumentation, the segmentation accuracy of our model on the Massachusetts roads dataset is 95.3%. Moreover, our models perform better than other models when training with the same loss function and experimental settings. The experiment result shows that the dilated convolution with vertical and horizontal kernels will enhance the neural network on linear feature extraction.

Parking Slot Detection on Around-View Images Using DCNN.

Due to the complex visual environment and incomplete display of parking slots on around-view images, vision-based parking slot detection is a major challenge. Previous studies in this field mostly use the existing models to solve the problem, the steps of which are cumbersome. In this paper, we propose a parking slot detection method that uses directional entrance line regression and classification based on a deep convolutional neural network (DCNN) to make it robust and simple. For parking slots with different shapes and observed from different angles, we represent the parking slot as a directional entrance line. Subsequently, we design a DCNN detector to simultaneously obtain the type, position, length, and direction of the entrance line. After that, the complete parking slot can be easily inferred using the detection results and prior geometric information. To verify our method, we conduct experiments on the public ps2.0 dataset and self-annotated parking slot dataset with 2,135 images. The results show that our method not only outperforms state-of-the-art competitors with a precision rate of 99.68% and a recall rate of 99.41% on the ps2.0 dataset but also performs a satisfying generalization on the self-annotated dataset. Moreover, it achieves a real-time detection speed of 13 ms per frame on Titan Xp. By converting the parking slot into a directional entrance line, the specially designed DCNN detector can quickly and effectively detect various types of parking slots.

[Chemome profiling and comparison of three Orobanche medicinal plants].

Several Orobanche medicinal plants sometimes served as alternative sources of Cistanches Herba, attributing to the benefits such as tonifying kidney, strengthening tendons and bones. Among them, O. coerulescens, O. cernua and O. pycnostachya have been widely utilized in northern China for treatments of pains in the loins and knees, impotence, and spermatorrhea. However, their chemical profiles haven't been elucidated. In the present study, UHPLC-IT-TOF-MS was implemented to conduct in-depth chemome profiling of O. coerulescens, O. cernua and O. pycnostachya, aiming to achieve a comprehensive chemical characterization and to provide pronounced information for the quality control and clinical applications. An ACE Ultra-Core 2.5 Super C_(18)(3.0 mm×150 mm, 2.5 µm) column was deployed for chromatographic separations, and high-resolution MS~n spectra were recorded by IT-TOF-MS. Forty-eight components, in total, were observed, and thirty-eight ones were structurally annotated according to proposing mass fragmentation patterns, matching with relevant databases. Particularly, nine ones were confirmed by reference compounds. Overall, the chemical compositions of O. coerulescens and O. cernua are quite similar, and differences occur between O. pycnostachya and the prior two ones; primary chemical family is phenylethanoid glycosides, and several lignan glycosides as well as iridoid glycosides are also observed; the primary components include acteoside, isoacteoside, crenatoside and 2'-acetylacteoside, etc.

Vacant Parking Slot Detection in the Around View Image Based on Deep Learning.

Due to the complex visual environment, such as lighting variations, shadows, and limitations of vision, the accuracy of vacant parking slot detection for the park assist system (PAS) with a standalone around view monitor (AVM) needs to be improved. To address this problem, we propose a vacant parking slot detection method based on deep learning, namely VPS-Net. VPS-Net converts the vacant parking slot detection into a two-step problem, including parking slot detection and occupancy classification. In the parking slot detection stage, we propose a parking slot detection method based on YOLOv3, which combines the classification of the parking slot with the localization of marking points so that various parking slots can be directly inferred using geometric cues. In the occupancy classification stage, we design a customized network whose size of convolution kernel and number of layers are adjusted according to the characteristics of the parking slot. Experiments show that VPS-Net can detect various vacant parking slots with a precision rate of 99.63% and a recall rate of 99.31% in the ps2.0 dataset, and has a satisfying generalizability in the PSV dataset. By introducing a multi-object detection network and a classification network, VPS-Net can detect various vacant parking slots robustly.