Search

Vaginitis is a prevalent gynecologic disease that threatens millions of women’s health. Although microscopic examination of vaginal discharge is an effective method to identify vaginal infections, manual analysis of microscopic leucorrhea images is extremely time-consuming and labor-intensive. To automate the detection and identification of visible components in microscopic leucorrhea images for early-stage diagnosis of vaginitis, we propose a novel end-to-end deep learning-based cells detection framework using attention-based detection with transformers (DETR) architecture. The transfer learning was applied to speed up the network convergence while maintaining the lowest annotation cost. To address the issue of detection performance degradation caused by class imbalance, the weighted sampler with on-the-fly data augmentation module was integrated into the detection pipeline. Additionally, the multi-head attention mechanism and the bipartite matching loss system of the DETR model perform well in identifying partially overlapping cells in real-time. With our proposed method, the pipeline achieved a mean average precision (mAP) of 86.00% and the average precision (AP) of epithelium, leukocyte, pyocyte, mildew, and erythrocyte was 96.76, 83.50, 74.20, 89.66, and 88.80%, respectively. The average test time for a microscopic leucorrhea image is approximately 72.3 ms. Currently, this cell detection method represents state-of-the-art performance.

Accompanied with the rapid increase of the demand for routine examination of leucorrhea, efficiency and accuracy become the primary task. However, in super depth of field (SDoF) system, the problem of automatic detection and localization of cells in leucorrhea micro-images is still a big challenge. The changing of the relative position between the cell center and focus plane of microscope lead to variable cell morphological structure in the two-dimensional image, which is an important reason for the low accuracy of current deep learning target detection algorithms. In this paper, an object detection method based on Retinanet in state of super depth of field is proposed, which can achieve high precision detecting of leucorrhea components by the SDoF feature aggregation module. Compared with the current mainstream algorithms, the mean average accuracy (mAP) index has been improved significantly, the mAP index is 82.7% for SDoF module and 83.0% for SDoF+ module, with an average increase of more than 10%. These improved features can significantly improve the efficiency and accuracy of the algorithm. The algorithm proposed in this paper can be integrated into the leucorrhea automatic detection system.

Here we proposed a single A-line sample-structure-nondependent (SSNd) dispersion detection and compensation method for Fourier-domain optical coherence tomography (FD-OCT), without need for acquiring and processing B-scan data. A new FD-OCT dispersion mismatch index, based on the line slope of the bright line in the A-line spectrogram, has been presented in the proposed method. With the new dispersion mismatch index, the proposed single A-line method can fast and visually detect the dispersion mismatch states of FD-OCT setup and perform the SSNd dispersion compensation, just using a single A-line. Experimental results of multiple samples demonstrated the advantages and convenience of the proposed method, and also proved that the proposed method can be used to analyze the relation of OCT imaging and optical path difference between the sample and reference arms.

The vestibular system is the sensory apparatus that helps the body maintain its postural equilibrium, and semicircular canal is an important organ of the vestibular system. The semicircular canals are three membranous tubes, each forming approximately two-thirds of a circle with a diameter of approximately 6.5 mm, and segmenting them accurately is of great benefit for auxiliary diagnosis, surgery, and treatment of vestibular disease. However, the semicircular canal has small volume, which accounts for less than 1% of the overall computed tomography image. Doctors have to annotate the image in a slice-by-slice manner, which is time-consuming and labor-intensive. To solve this problem, we propose a novel 3D convolutional neural network based on 3D U-Net to automatically segment the semicircular canal. We added the spatial attention mechanism of 3D spatial squeeze and excitation modules, as well as channel attention mechanism of 3D global attention upsample modules to improve the network performance. Our network achieved an average dice coefficient of 92.5% on the test dataset, which shows competitive performance in semicircular canals segmentation task.

Phase unwrapping is a critical step to obtaining a continuous phase distribution in optical phase measurements and coherent imaging techniques. Traditional phase-unwrapping methods are generally low performance due to significant noise or undersampling. This paper proposes a deep convolutional neural network (DCNN) with a weighted jump-edge attention mechanism, namely, VDE-Net, to realize effective and robust phase unwrapping. Experimental results revealed that the weighted jump-edge attention mechanism, which is first proposed and simple to calculate, is useful for phase unwrapping. The proposed algorithm outperformed other networks or common attention mechanisms. In addition, an unseen wrapped phase image of a living red blood cell (RBC) was successfully unwrapped by the trained VDE-Net, thereby demonstrating its strong generalization capability.

Accompanied with the clinical routine examination demand increase sharply, the efficiency and accuracy is the first priority. However, the issues on the automatic detection and localization of cells in microscopic images of specific biological sample in super-depth-of-field (SDoF) system remains great challenges. We proposed object detection method for cells and other phenotypic components in the SDoF micrograph based on Retinanet. Compared the current mainstream algorithm, the mean average precision (mAP) index is significantly improved. In the experiment of leucorrhea samples and fecal samples, mAP indexes are 83.1% and 88.1% respectively, with an average increase of 10%. These improved characteristics can put the efficiency and accuracy forward significantly. The algorithm proposed in this paper can be integrated into the fecal and leucorrhea automatic detection system.

Vaginitis is a gynecological disease affecting the health of millions of women all over the world. The traditional diagnosis of vaginitis is based on manual microscopy, which is time-consuming and tedious. The deep learning method offers a fast and reliable solution for an automatic early diagnosis of vaginitis. However, deep neural networks require massive well-annotated data. Manual annotation of microscopic images is highly cost extensive because it not only is a time-consuming process but also needs highly trained people (doctors, pathologists, or technicians). Most existing active learning approaches are not applicable in microscopic images due to the nature of complex backgrounds and numerous formed elements. To address the problem of high cost of labeling microscopic images, we present a data-efficient framework for the identification of vaginitis based on transfer learning and active learning strategies. The proposed informative sample selection strategy selected the minimal training subset, and then the pretrained convolutional neural network (CNN) was fine-tuned on the selected subset. The experiment results show that the proposed pipeline can save 37.5% annotation cost while maintaining competitive performance. The proposed promising novel framework can significantly save the annotation cost and has the potential of extending widely to other microscopic imaging applications, such as blood microscopic image analysis.

The avalanche photodiode (APD) chip is the core component of the transistor outline (TO). The concentricity between the inner circle (IC) of the APD active area and the outer circle (OC) of the TO base will directly affect a component’s key performance indicators, such as external quantum efficiency, receiving sensitivity and responsivity, thereby impacting quality assurance, performance improvement, and stable operation. Nevertheless, as the surge in demand for components increases, the traditional visual inspection relying on manual and microscope has been unable to meet the requirements of mass manufacturing for real-time quality and efficiency. Thus, a Concentricity Microscopic Vision Measurement System (CMVMS) mainly composed of a microscopic vision acquisition unit and an intelligent concentricity measurement unit has been proposed, designed, and implemented. On the basis of analyzing the 3D complex environment of TO components, a coaxial illumination image acquisition scheme that can take into account the characteristics of the OC and IC has been proposed. Additionally, a concentricity image measurement method based on dynamic threshold segmentation has been designed to reduce the interference of complex industrial environment changes on measurement accuracy. The experiment results show that the measurement accuracy of the CMVMS system is over 97%, and with a single measurement time of less than 0.2s, it can better meet the real-time and accuracy requirements. To the best of our knowledge, this is the first report on the realization of real-time concentricity measurement in optical component packaging, and this technology can be extended to other fields of concentricity measurement.

Microscopic laser engraving surface defect classification plays an important role in the industrial quality inspection field. The key challenges of accurate surface defect classification are the complete description of the defect and the correct distinction into categories in the feature space. Traditional classification methods focus on the terms of feature extraction and independent classification, therefore, feed handcrafted features may result in useful feature loss. In recent years, convolutional neural networks (CNNs) have achieved excellent results in image classification tasks with the development of deep learning. Deep convolutional networks integrate feature extraction and classification into self-learning, but require large datasets. The training datasets for microscopic laser engraving image classification are small, therefore, we used pre-trained CNN models and applied two fine-tuning strategies. Transfer learning proved to perform well even on small future datasets. The proposed method was evaluated on the datasets consisting of 1986 laser engraving images captured by a metallographic microscope and annotated by experienced staff. Because handcrafted features were not used, our method is more robust and achieves better results than traditional classification methods. Under five-fold-validation, the average accuracy of the best model based on DenseNet121 is 96.72%.

Bump offsets have harmful effects on the quality of flex-on-glass (FOG) bonding in liquid crystal display (LCD) module industry. Here we report a novel automatic optical inspection (AOI) method to achieve fast inspecting FOG bonding bump offsets. Overcoming difficulties to image targets at different depths in light absorption medium when using ultra-short-exposure and high-speed imaging, the proposed method innovatively achieves inspecting FOG bonding bump offsets using differential interference contrast imaging based on spatial Poisson point process approximation. Theoretical analysis proves that the proposed method has high reliability in FOG bonding bump offset AOI inspection. Experimental results show missed detection rate of the inspection algorithm is 0.01%, and false rate is 0.1%, and verify the accuracy and validity of the proposed method, all of which are significant for guiding practice in LCD module industry.

Background: Xanthium sibiricum is a well-known traditional Chinese medicine (TCM) that has been commonly used to treat rhinitis and related nasal diseases. The aim of this study was to develop a comprehensive analytical method based on high-performance liquid chromatographyelectrospray ionization coupled with triple quadrupole-linear ion trap mass spectrometry (LC-ESIQTRAP- MS/MS) for the simultaneous determination of phenolic acids, anthraquinones, and flavonoids in the aerial part and fruit of Xanthium sibiricum. Methods: The separation was completed on Agilent ZORBAX SB-C18 column (250 × 4.6 mm, 5μm) using methanol and 0.2% (v/v) aqueous formic acid as the mobile phase. The target components were analyzed in negative ion mode with accurate and sensitive multiple reaction monitoring (MRM) mode. Results: The correlation coefficients of all the calibration curves were higher than 0.9994. Relative standard deviations of intra- and inter-day precisions of the eighteen components were all lower than 2.87% and the recoveries were in the range from 97.73% to 101.82%. The validated method was successfully applied to possess forty Xanthium sibiricum samples (Xanthii Herba, Xanthii Fructus, and processed Xanthii Fructus) collected from different places in P. R. China. Furthermore, principal component analysis (PCA) was performed to evaluate and classify the samples according to the contents of the eighteen bioactive components. Conclusion: All the results demonstrated that the developed method was useful and could be applied for the overall assessment of the quality of Xanthii Herba and Xanthii Fructus.

Background: Pseudostellaria heterophylla is an important tonic traditional Chinese medicine. However, the molecular changes in the herb from geo-authentic habitat and cultivated bases remain to be explored. Objective: The purpose of this research was to study differences in P. heterophylla from geo-authentic habitat and cultivated bases. Methods: High-throughput technologies of transcriptomic and proteomic were used to identify proteins. Isobaric Tags for Relative and Absolute Quantification (iTRAQ) MS/MS has been utilized to evaluate changes in P. heterophylla from geo-authentic habitat and cultivated bases. Results: In this study, a total of 3775 proteins were detected, and 140 differentially expressed proteins were found in P. heterophylla from geo-authentic habitat and cultivated bases. 44 significantly differential expressed proteins were identified based on functional analysis classified into nine categories. Five differentially expressed proteins were confirmed at the gene expression level by Quantitative realtime PCR. Catabolic metabolism, carbohydrate metabolism, and response to stress of oxidoreductases and transferases in P. heterophylla from geo-authentic habitat were stronger than in those from cultivated bases, but protein folding and response to stress of heat shock proteins, isomerases, rubisco large subunit-binding proteins, chaperone proteins, and luminal-binding proteins in herbs from cultivated bases were more active. ADG1 and TKTA could be the critical proteins to regulate sucrose; MFP2 and CYS may be the crucial proteins that control the metabolism of fatty acids and amino acids. Conclusion: These results will provide the basic information for exploring the differences in secondary metabolites in P. heterophylla from geo-authentic habitat and cultivated bases and the protein mechanism of its quality formation.

Trichomonas examination is one of the important items in the leucorrhea routine detection. And it cannot be recognized by still images because of the unstable morphology and unfixed focal location caused by motion characteristic. We proposed an improved VIBE algorithm. 6 videos (totally 1414 frames) are collected for testing. In order to compare the effects of the algorithms, we segment each frame artificially as ground truth. Experiments show that percentage of correct classification (PCC) achieves 88%. The proposed improved method can effectively suppress the false detection caused by the formed components such as epithelial cells in the leucorrhea microscopic image and the missed detection caused by the background model update during the movement. At the same time, improvements can effectively suppress smear and ghost areas. The algorithm proposed in this paper can be integrated into the leucorrhea automatic detection system.

Mobile manipulator could perform tasks in a larger workspace than fixed based manipulators. However, it also brings challenges to high precision kinematic control. The difficulties are mainly concentrated in the system redundancy and sensor noises. In this paper, we propose a novel kinematic controller for a kind of mobile manipulators, which is capable of handling system redundancy, physical constraints and noises. The controller is designed in velocity level, and is compliance to joint angles, velocities as well as accelerations.

Diagnosis of Trichomonas vaginalis infection is one of the most important factors in the routine examination of leucorrhea. According to the motion characteristics of Trichomonas vaginalis, a viable detection method is the use of a microscopic camera to record videos of leucorrhea samples and video object detection algorithms for detection. Most Trichomonas vaginalis is defocused and displays as shadow regions on microscopic images, and it is hard to recognize the movement of shadow regions using traditional video object detection algorithms. In order to solve this problem, we propose two convolutional neural networks based on an encoder-decoder architecture. The first network has the ability to learn the difference between frames and utilizes the image and optical flow information of three consecutive frames as the input to perform rough detection. The second network corrects the coarse contours and uses the image information and the rough detection result of the current frame as the input to perform fine detection. With these two networks applied, the metric value of the mean intersection over union of Trichomonas vaginalis achieves 72.09% on test videos. The proposed networks can effectively detect defocused Trichomonas vaginalis and suppress false alarms caused by the motion of formed elements or impurities.

Fecal samples can easily be collected and are representative of a person’s current health state; therefore, the demand for routine fecal examination has increased sharply. However, manual operation may pollute the samples, and low efficiency limits the general examination speed; therefore, automatic analysis is needed. Nevertheless, recognition exhaust time and accuracy remain major challenges in automatic testing. Here, we introduce a fast and efficient cell detection algorithm based on the Faster-R-CNN technique and is the Resnet-152 convolutional neural network architecture. Additionally, a region proposal network and a network combined with principal component analysis were proposed for cell location and recognition in microscopic images. Our algorithm achieved a mean average precision of 84% and a 723 ms detection time per sample for 40560 fecal images. Thus, this approach may provide a solid theoretical basis for real-time detection in routine clinical examinations while accelerating the process to satisfy increasing demand.

Balanced dispersion between reference and sample arms is critical in frequency-domain optical coherence tomography (FD-OCT) to perform imaging with the optimal axial resolution, and the spectroscopic analysis of each voxel in FD-OCT can provide the metric of the spectrogram. Here we revisited dispersion mismatch in the spectrogram view using the spectroscopic analysis of voxels in FD-OCT and uncovered that the dispersion mismatch disturbs the A-scan’s spectrogram and reshapes the depth-resolved spectra in the spectrogram. Based on this spectroscopic effect of dispersion mismatch on A-scan’s spectrogram, we proposed a numerical method to detect dispersion mismatch and perform dispersion compensation for FD-OCT. The proposed method can visually and quantitatively detect and compensate for dispersion mismatch in FD-OCT, with visualization, high sensitivity, and independence from sample structures. Experimental results of tape and mouse eye suggest that this technique can be an effective method for the detection and compensation of dispersion mismatch in FD-OCT.

Obtaining microscopic images of all cells in a specimen with thick liquid layers as clearly and accurately as possible is essential for the diagnosis of disease or infection in patients using an automated in vitro diagnostics (IVD) analyzer. However, the autofocus method used by traditional analyzers can only clearly image cells located on the same horizontal plane. For cells distributed on different horizontal planes in a thick liquid layer, it is challenging to image all of them clearly. In this paper, combining technologies such as autofocus, dynamic fusion, and adaptive trajectory planning, we present and validate a novel system to achieve the rapid localization and accurate imaging of all cells in thick liquid layers. Through the Non-False MultiFocus image fusion algorithm, the clear cell images located in different liquid layers are fused together. The adaptive path planning is further used to shorten the multi-field autofocus time and improve the analyzer’s working efficiency. This system includes a microscope, CMOS camera, light grating, stepping motor, motion control circuit, and other hardware modules, which are controlled by a software algorithm. The feces sediment analyzer has been successfully deployed to the pathology departments of multiple hospitals.

Skin plays a significant role in human body function and its collagen states change during the human skin ageing process, which affects skin function. We previously reported on geometry-dependent spectroscopic contrast achieved by spectroscopic micro-optical coherence tomography (SμOCT), which discovered that transversely oriented and regularly arranged nano-cylinders selectively backscatter the long-wavelength lights and generate spectral centroid (SC) shifts towards the long wavelengths within a spectral window of 700 - 950 nm. Here we further proposed a novel method towards indicating the state of human skin in vivo using geometry-dependent spectroscopic contrast imaging. The proposed method can obtain spectroscopic contrast images of different human skin layers, including the papillary dermis and reticular dermis, and provide a quantitative method towards indicating collagen state in human skin in vivo. Experimental results have shown that the proposed method may potentially have the ability to assess the state of human skin. Agency for Science, Technology and Research (A*STAR) Ministry of Education (MOE) Ministry of Health (MOH) This research was supported by the Singapore Ministry of Health’s National Medical Research Council under its Open Fund Individual Research Grant (MOH-OFIRG19may-0009), Ministry of Education Singapore under its Academic Research Fund Tier 1 (2018-T1-001-144), Agency for Science, Technology and Research (A*STAR) under its Industrial Alignment Fund (Pre-positioning) (H17/01/a0/008) and the National Natural Science Foundation of China under Grant 61905036.

Purpose Detection of leukocytes is critical for the routine leukorrhea exam, which is widely used in gynecological examinations. An elevated vaginal leukocyte count in women with bacterial vaginosis is a strong predictor of vaginal or cervical infections. In the routine leukorrhea exam, the counting of leukocytes is primarily performed by manual techniques. However, the viewing and counting of leukocytes from multiple high-power viewing fields on a glass slide under a microscope leads to subjectivity, low efficiency, and low accuracy. To date, many biological cells in stool, blood, and breast cancer have been studied to realize computerized detection; however, the detection of leukocytes in microscopic leukorrhea images has not been studied. Thus, there is an increasing need for computerized detection of leukocytes. Methods There are two key processes in the computerized detection of leukocytes in digital image processing. One is segmentation; the other is intelligent classification. In this paper, we propose a combined ensemble to detect leukocytes in the microscopic leukorrhea image. After image segmentation and selecting likely leukocyte sub-images, we obtain the leukocyte candidates. Then, for intelligent classification, we adopt two methods: feature extraction and classification by a support vector machine (SVM); applying a modified convolutional neural network (CNN) to the larger sub-images. If different methods classify a candidate in the same category, the process is finished. If not, the outputs of the methods are provided to a classifier to further classify the candidate. Results After acquiring leukocyte candidates, we attempted three methods to perform classification. The first approach using features and SVM achieved 88% sensitivity, 97% specificity, and 92.5% accuracy. The second method using CNN achieved 95% sensitivity, 84% specificity, and 89.5% accuracy. Then, in the combination approach, we achieved 92% sensitivity, 95% specificity, and 93.5% accuracy. Finally, the images with marked and counted leukocytes were obtained. Conclusion A novel computerized detection system was developed for automated detection of leukocytes in microscopic images. Different methods resulted in comparable overall qualities by enabling computerized detection of leukocytes. The proposed approach further improved the performance. This preliminary study proves the feasibility of computerized detection of leukocytes in clinical use. This article is protected by copyright. All rights reserved.

Scrophulariae Radix is one of the most popular traditional Chinese medicines. The harvesting time of Scrophulariae Radix is closely related to the quality of products in this traditional Chinese medicine. The goal of the study is to analyze the dynamic changes of metabolite accumulation in Scrophulariae Radix. The difference of constituents in Scrophulariae Radix harvested at different times was analyzed by liquid chromatography with triple quadrupole-time of flight tandem mass spectrometry coupled with principal component analysis and partial least-square discriminant analysis. According to the accurate mass of molecular and product ions provided by triple quadrupole-time of flight tandem mass spectrometry, a total of 30 differential constituents were identified. Furthermore, the contents of ten index differential constituents in Scrophulariae Radix were simultaneously determined by liquid chromatography coupled with triple quadrupole-linear ion trap tandem mass spectrometry. Gray relational analysis was performed to evaluate the samples harvested at different times according to the contents of ten constituents. All of the results demonstrated that the quality of Scrophulariae Radix collected at traditional harvest time was better. This study will provide the basic information for revealing the dynamic change law of metabolite accumulation of Scrophulariae Radix and exploring its quality forming mechanism.

The pressing deformation of conductive microspheres in Anisotropic Conductive Film(ACF) bonding mainly decides the interconnection resistance and bonding quality. Prior works have studied the formation mechanism of microspheres’ pressing deformation and bonding resistance, but haven't theorized how to obtain them. This paper has propose to accurately inspect the pressing deformation by differential interference contrast(DIC) imaging and images analyzing. By the deformation inspection, it further proposed to calculate the bonding resistance with the resistance model. Experimental results show it can accurately inspect the conductive microspheres’ pressing deformation and interconnection resistance, showing that the unacceptable bumps have a big resistance, over 300 mΩ, while the acceptable bumps have a small resistance, about 126 mΩ. The inspection result is more consistent with the actual resistance than ones from the references for it doesn't have the shortages of parameters acquisition in the references. This paper has theorized how to inspect the ACF bonding resistance which will be useful to electronic packaging, especially in LCD module industry.

Leucorrhea routine test is one of the most widely used tests in gynecological examinations, and fungi inspection is vital for gynecological test because fungi is an important evidence for fungal vaginitis. In order to improve detection accuracy, an automatic identification of fungi in microscopic leucorrhea images based on convolutional neural network (CNN) and morphological method is proposed in this paper. First, we use the maximum inter-class variance method to segment original image and obtain possible fungi subimages. Then, a fully trained CNN is applied to recognize fungi. Finally, morphological method, such as template match method and concave point detection method, is used to further classify the selected candidate to improve recognize accuracy. In experiments, the method using CNN and morphological method achieved 93.26% accuracy.

Fecal microscopic examination is a routine examination item to determine whether the digestive system is normal by analyzing formed elements. Traditional method is that doctor uses microscope eyepiece to observe sample smears. The efficiency is low, and examination results depend on doctor's experience level. Therefore, intelligent identification of formed elements is the main development direction of current fully automated fecal instruments. Unlike blood or urine samples, human fecal samples contain a lot of impurities, and sample stratification phenomenon is serious. So image quality assessment methods are difficult to find the sharpest image, affecting effectiveness of intelligent identification algorithm. In this paper, the microscopic image autofocus technology for human fecal samples is studied and divided into two parts: location and photographing. In location process, we use SMD algorithm to determine sample photographing interval. In photographing process, microscope platform zigzagged move in the interval to obtain each view's successively image sequences of different focal lengths. In order to accurately find the sharpest image in image sequence, we compared the difference between human eyes with 31 types of no-reference image quality assessment methods based on entropy, gradient, color, edge, contrast, similarity, and transform domain. Finally an improved Local TV algorithm was chose. Experimental results show that the improved Local TV algorithm is insensitive to changes in sample concentration with good robustness, and the accuracy rate can reach 94.26%. Our experimental results have some reference value for other focusing problems of complex microscopic images.

Pseudostellariae Radix is an important traditional Chinese medicine (TCM), which is consumed commonly for its positive health effects. However, a lack of transcriptomic and genomic information hinders research on Pseudostellariae Radix. Here, high-throughput RNA sequencing (RNA-seq) was employed for the de novo assembly to analyze the transcriptome in Pseudostellariae Radix, finding significantly differentially expressed genes in this TCM from different fields based on RNA-seq and bioinformatic analysis. A total of 146,408,539 paired-end reads were generated and assembled into 89,857 unigenes with an average length of 862bp. All of the assembly unigenes were annotated by running BLASTx and BLASTn similarity searches on the Non-redundant nucleotide database (NT), the Non-redundant protein database (NR), Swiss-Prot, Cluster of Orthologous Groups (COG), Kyoto Encyclopedia of Genes and Genomes (KEGG), Gene Ontology (GO), and Interpro. On the basis of bioinformatic analysis and the expression profiles for Pseudostellariae Radix, 29 significantly differentially expressed genes were identified, which provides the basic information for exploring the molecular mechanisms that determine the quality of Pseudostellariae Radix from different fields. The expression levels of 29 genes were validated by real-time quantitative PCR (RT-qPCR). This is the first study to sample Pseudostellariae Radix, which provides an invaluable resource for understanding the genome of this herb.

The analysis of fecal-type components for clinical diagnosis is important. The main examination involves the counting of red blood cells (RBCs), white blood cells (WBCs), and molds under the microscopic. With the development of machine vision, some vision-based detection schemes have been proposed. However, these methods have a single target for detection, with low detection efficiency and low accuracy. We proposed an algorithm to identify the visible image of fecal composition based on intelligent deep learning. The algorithm mainly includes region proposal and candidate recognition. In the process of segmentation, we proposed a morphology extraction algorithm in a complex background. As for the candidate recognition, we proposed a new convolutional neural network (CNN) architecture based on Inception-v3 and principal component analysis (PCA). This method achieves high-average Precision of 90.7%, which is better than the other mainstream CNN models. Finally, the images within the rectangle marks were obtained. The total time for detection of an image was roughly 1200 ms. The algorithm proposed in the present paper can be integrated into an automatic fecal detection system.

In order to solve the problem of high false detection ration and poor real-time performance in the automated and optical inspection of surface defect on cartridge case, a visually high-speed method was proposed. Taking a typical 9mm cartridge case of handgun as the object and considering its surface material, types of defect and detection method, an optical detection system was established on the basis of machine vision. Through studies of the metal texture on the surface, the detection operator called LBPV which was combined local binary pattern with local image variance was proposed. On the basis of the parallel architecture and hardware characteristic of GPU, texture memory and shared memory was introduced to reduce the complexity of this algorithm. Experimental results show that the algorithm can effectively detect and recognize the surface defect on cartridge case. Accuracy reaches up to 95.3% and missing rate is 0%. Real-time performance was improved as a result of GPU acceleration for LBPV operator, the acceleration rate is about 60x. Real-time detection with high accuracy in industrial production were satisfied.

Pseudostellariae Radix (PR) is an important traditional Chinese medicine (TCM), which is consumed commonly for its positive health effects. However, the chemical differences of PR from different cultivated fields and germplasms are still unknown. In order to comprehensively compare the chemical compositions of PR from different cultivated fields, in this study, ¹H-NMR-based metabolomics coupled with high performance liquid chromatography (HPLC) were used to investigate the different metabolites in PR from five germplasms (jr, zs1, zs2, sb, and xc) cultivated in traditional fields (Jurong, Jiangsu, JSJR) and cultivated fields (Zherong, Fujian, FJZR). A total of 34 metabolites were identified based on ¹H-NMR data, and fourteen of them were found to be different in PR from JSJR and FJZR. The relative contents of alanine, lactate, lysine, taurine, sucrose, tyrosine, linolenic acid, γ-aminobutyrate, and hyperoside in PR from JSJR were higher than that in PR from FJZR, while PR from FJZR contained higher levels of glutamine, raffinose, xylose, unsaturated fatty acid, and formic acid. The contents of Heterophyllin A and Heterophyllin B were higher in PR from FJZR. This study will provide the basic information for exploring the influence law of ecological environment and germplasm genetic variation on metabolite biosynthesis of PR and its quality formation mechanism.

Scrophulariae Radix is one of the most popular traditional Chinese medicines (TCMs), which needs to be processed by ‘sweating’ methods. Primary processing of Scrophulariae Radix is an important link which closely relates to the quality of products in this TCM. To facilitate selection of the suitable ‘sweating’ processing method for Scrophulariae Radix, in this study the quality of Scrophulariae Radix processed by different ‘sweating’ methods was evaluated based on simultaneous determination of multiple bioactive constituents combined with grey relational analysis. The contents of iridoid glycosides, phenylpropanoid glycosides, and organic acids in Scrophulariae Radix processed by different ‘sweating’ methods were simultaneously determined using ultra high performance liquid chromatography coupled with triple quadrupole-linear ion trap mass spectrometry (UPLC-QTRAP-MS/MS). Furthermore, grey relational analysis (GRA) was performed to evaluate the ‘sweating’ processed samples according to the contents of twelve constituents. All of the results demonstrated that the quality of Scrophulariae Radix processed by oven drying at 35 °C and ‘sweating’ for three days was better. The developed method was useful for the overall assessment on quality of Scrophulariae Radix, and this study may provide the foundation and support for ‘sweating’ processing of Scrophulariae Radix in normalization and standardization.

A nanometric surface plasmon polariton (SPP) filter based on a vertical coupled metal–insulator–metal (MIM) Fabry–Perot resonator is proposed and analyzed. The transmission characteristics of the SPP filter are analyzed in detail by using the finite difference time domain method. And the resonance condition derived by the numerical method is consonant with the analytic model based on the Fabry–Perot model, which includes the metal loss and dispersion effects. And the simulation results show that multiple transmission dips can be formed and the resonance wavelengths of the transmission dips can be altered by the Fabry–Perot cavity length and width. Also the extinction ratio and the quality factor of the filter are affected by the barrier thickness of the coupling region. The proposed nanometric plasmonic filter is simple and very promising for the SPP waveguides platform.

A comprehensive analytical method based on UPLC-MS/MS was developed for the simultaneous determination of thirteen components including three stilbenes (stilbeneglucoside, polydatin, resveratrol), four anthraquinones (emodin, physcion, emodin-8-β-D-glucopyranoside, aloe-emodin), five flavonoids (epicatechin, rutin, hyperoside, astragalin,quercetin) and one phenolic acid (gallic acid) in Polygoni Multifori Caulis.The separation was carried out on a Waters BEH C18 column（2.1 mm×100 mm,1.7 μm）with gradient elution of acetonitrile-water (0.1% acetic acid) at a flow rate of 0.25 mL•min⁻¹, and column temperature was 35 ℃. The target compounds were analyzed by multiple reaction monitoring (MRM) mode. TOPSIS analysis ware performed to evaluate the samples from different areas and commercial herbs according to the contents of thirteen components. The correlation coefficients of all the calibration curves were higher than 0.991 5. The average recoveries ranged from 95.24% to 102.3%, and the relative standard deviations were less than 5%. The result of TOPSIS analysis showed that the comprehensive quality of Polygoni Multifori Caulis sample from Guangzhou was better. The developed method with good repeatability and accuracy was suitable for the simultaneous determination of multiple functional substances, which provided a new basis for the comprehensive assessment and overall control of the quality of Polygoni Multifori Caulis.

Pseudostellaria heterophylla is an important traditional Chinese herbal medicine with vast clinical consumption because of its positive effects. To date, changes in the metabolite composition of P. heterophylla have been well documented; however, the molecular differences between cultivated P. heterophylla and its wild-type are still unknown. The aim of this study was to investigate differences in cultivated and wild P. heterophylla. Due to the lack of a genomic database, we used high-throughput transcriptomic and proteomic technologies to identify proteins in the herb. Isobaric tags for relative and absolute quantification (iTRAQ) MS/MS were used to detect statistically significant changes between cultivated and wild P. heterophylla. We detected 3775 proteins; 332 showed differential accumulations across two different ecotypes of P. heterophylla. 71 significant differential expressions of proteins were selected based on GO annotations, KEGG, STRING analysis, and expression level. Quantitative real-time PCR analysis confirmed the authenticity and accuracy of the proteomic analysis. The results indicated that the carbohydrate and cellular amino acid metabolisms in cultivated P. heterophylla were weaker than those in its wild-type; seven important proteins were found to regulate sucrose and amino acids. This will provide the basic information for exploring the cause of differences in secondary metabolites in different ecotypes of P. heterophylla and the protein mechanism of its quality formation. Biological significance This study combined a transcriptome, proteome, and metabolism approach for analyzing differentially expressed proteins of cultivated and wild P. heterophylla and established the relationship between significantly differentially expressed proteins and differential chemical components in non-model plants. The results of proteomic analysis provide the basic information for exploring the quality forming process, which will demonstrate, and provide guidance for, the study of effective constituents of P. heterophylla and its quality formation mechanism.