Discovery of potential anti- Staphylococcus aureus natural products and their mechanistic studies using machine learning and molecular dynamic simulations.

The structure-activity analysis (SAR) and machine learning were used to investigate potential anti-S. aureus agents in a faster method. In this study, 24 oxygenated benzene ring components with S. aureus inhibition capacity were confirmed by literature exploring and in-house experiments, and the SAR analysis suggested that the hydroxyl group position may affect the anti-S. aureus activity. The 2D-MLR-QSAR model with 9 descriptors was further evaluated as the best model among the 21 models. After that, hesperetic acid and 2-HTPA were further explored and evaluated as the potential anti-S. aureus agents screening in the natural product clustering library through the best QSAR model calculation. The antibacterial capacities of hesperetic acid and 2-HTPA had been investigated and proved the similar predictive pMIC value resulting from the QSAR model. Besides, the two novel components were able to inhibit the growth of S. aureus by disrupting the cell membrane through the molecular dynamics simulation (MD), which further evidenced by scanning electron microscopy (SEM) test and PI dye results. Overall, these results are highly suggested that QSAR can be used to predict the antibacterial agents targeting S. aureus, which provides a new paradigm to research the molecular structure-antibacterial capacity relationship.

Spatio-temporal evolution and driving factors of green innovation efficiency in the Chinese urban tourism industry based on spatial Markov chain.

Green innovation in the tourism industry is a sustainable development concept for resource conservation and environmental optimization. The effective measurement of green innovation efficiency in the tourism industry and an accurate understanding of its spatial relationship was significantly important for promoting its sustainable development. Using the SBM-undesirable model, kernel density estimation, and a spatial Markov chain, we explored the spatio-temporal evolution characteristics and influencing mechanisms of urban tourism green innovation efficiency (TGIE) in China between 2000 and 2020. We found that (1) the temporal and spatial changes of TGIE were generally at a lower than medium level and fluctuated throughout country, with a transition in the east, collapse in the middle, and stagnation in the northeast. (2) The dynamic evolution of TGIE always exhibited polarization, but regional coordination was gradually enhanced with strong stability, although it was difficult to achieve leap-forward development. The cities with spatial upward transfer were concentrated mainly in the central and western region and while there were few cities with a downward adjustment, there were obvious asymmetrical spatial spillover effects. (3) The driving factors of TGIE were the overall economic level, industrial structure, government regulation, and education level. These factors had a significant positive relationship with TGIE, while the degree of opening up to the outside world has no significant effect, but the degree of influence, mechanism, and conditions of each factor were strongly regional.

Fabrication of chitosan-based fluorescent hydrogel membranes cross-linked with bisbenzaldehyde for efficient selective detection and adsorption of Fe2.

Chitosan, as a natural biomass material, is green, recyclable, sustainable and well biocompatible. The molecular chain is rich in active groups such as amino and hydroxyl groups, and its preparation of fluorescent probes has the advantages of biocompatibility and efficient detection performance. In this study, a bis(benzaldehyde) (BHD) fluorescent functional molecule was designed. Then a series of fluorescent chitosan-based hydrogel films (CSBHD) were prepared using chitosan as raw material and BHD as cross-linking agent. As a fluorescent probe for metal ions, CSBHD was able to efficiently detect Fe2+ with a linear correlation of fluorescence intensity in the range of 0-160 µM, and the limit of detection (LOD) was 0.55 µM. Moreover, it has excellent adsorption performance for Fe2+ ions, with a maximum adsorption capacity of 223.5 g/mg at 500 mg/L Fe2+ concentration. Finally, we characterised the structure and microscopic morphology of CSBHD films and found that CSBHD as a hydrogel film has a high cross-linking density, good water resistance, excellent thermal stability, strong resistance to swelling, and excellent stability in cycling tests. Hence, it has great potential for application in adsorption and detection of Fe2+ ions. It also provides a good strategy for the application of chitosan based fluorescent probe materials in environmental monitoring and heavy metal ion adsorption.

A WeChat-based nursing intervention program improves the postoperative rehabilitation of breast cancer patients: results from a randomized controlled trial.

Background: In postoperative setting, breast cancer (BC) patients can experience adverse effects, including fatigue, sleep disorders, and pain, which substantially affect their health-related quality of life (HRQoL). This study sought to assess the effectiveness of a WeChat-based multimodal nursing program (WCBMNP) that was specifically designed for the rehabilitation of women following BC surgery. Methods: BC patients were randomly, single-blinded allocated to either the intervention (n=62) or control (n=63) cohorts. Over a period of 6 months (24 weeks), the intervention cohort received a WCBMNP in addition to routine nursing care, while the control cohort received routine nursing care only. To evaluate patients' fear of cancer recurrence (FCR), their overall fear score was assessed using the Japanese version of the Concerns About Recurrence Scale (CARS-J) for primary outcome. The initial outcome (HRQoL) and secondary results, such as fatigue, sleep, and pain, were examined using the Functional Assessment of Cancer Therapy-Breast (FACT-B, version 4.0) and Nursing Rating Scale (NRS), respectively. Results: Two hundred and ten participants, 85 participants were excluded. Compared to the controls (n=63), the intervention cohort (n=62) showed statistically significant improvements in their CARS-J scores. The intervention cohort aggregate scores on the FACT-B improved significantly but were affected by the compounding influences of cohort dynamics, temporal progression, and their interaction. Similar improvements were observed in the social/family and functional well-being domains. Emotional well-being was improved based on the effects of time and group-time interaction. In the intervention cohort, the "BC-specific subscale for additional concerns" was affected by group and time, whereas physical well-being was only affected by time. Conversely, there were no statistically significant changes in the variables of fatigue, sleep, and pain. Conclusions: The WCBMNP reduced FCR and significantly increased the HRQoL of female patients with BC postoperatively. The WCBMNP could be implemented as a postoperative rehabilitation intervention in this patient population to improve outcomes. Trial Registration: Chinese Clinical Trial Registry (ChiCTR2400081557).

Construction and evolutionary factors of spatial correlation network of China's provincial tourism resource conversion efficiency.

Research objective: To explore the spatial correlation network characteristics and formation mechanisms of tourism resource conversion efficiency, and provide reference for the collaborative improvement of tourism resource conversion efficiency at the provincial level in China. Research methods: Non parametric SBM efficiency measurement method and social network analysis method. Research hypothesis: The spatial network correlation characteristics of tourism resource conversion efficiency are obvious, and regional connections are close. Research findings: (i) during the research period, the spatial connection strength of China's tourism resource conversion efficiency continued to increase and the spatial network structure of tourism resource conversion efficiency tended to become more complex and significantly more stable. (ii) A spatially linked network with a stable tourism resource conversion efficiency structure formed in China. The number of network relations and density of the network fluctuated and increased, while the network efficiency continued to decrease; however, a strong small-world nature was observed. (iii) The economic development level difference matrix, tourism industry agglomeration difference matrix, human capital difference matrix, and marketization degree difference matrix significantly and positively affected spatial association relationship establishment, while the provincial adjacency matrix significantly and negatively affected such relationships.

Effects of sleep deprivation and hazard types on the hazard perception of young novice drivers: An ERP study.

OBJECTIVE: The present study aimed to explore the effects of sleep deprivation on young novice drivers' cognitive neural processing of different hazard types. METHOD: A 2 (sleep deprivation group, control group) × 3 (no hazard, covert hazard, overt hazard) mixed experimental design was used. Twenty-eight young drivers were sleep-deprived (no sleep within the past 24 h), while 28 drivers were in the control group (maintaining a normal schedule throughout the week). Eighty pictures containing a covert hazard (20 pictures), overt hazard (20 pictures) and no hazard (40 pictures) were presented. Participants were asked to press the keyboard quickly if they detected a hazard situation. The reaction time, accuracy, and changes in the N1 (100-150 ms) and N2 (250-350 ms) components of event-related potentials (ERP) measured using electroencephalography (EEG) were obtained. RESULTS: Compared to the control group, the response accuracy of sleep-deprived drivers was higher in the cover-hazard situation and their N1 latency was longer in the no-hazard situation. Compared to the no-hazard and overt-hazard situations, the participants' reaction times and N2 amplitudes were significantly greater, and the response accuracy was significantly lower in the covert-hazard situation. CONCLUSION: Hazard perception is compromised when drivers are sleep-deprived, especially when they are confronted with covert hazard situations. The findings help understand the negative effects of sleep deprivation in the early stage of young novice drivers' hazard perception.

Dominance complementation of parental heading date alleles of Hd1, Ghd7, DTH8, and PRR37 confers transgressive late maturation in hybrid rice.

Rice (Oryza sativa L.) is a short-day plant whose heading date is largely determined by photoperiod sensitivity (PS). Many parental lines used in hybrid rice breeding have weak PS, but their F1 progenies have strong PS and exhibit an undesirable transgressive late-maturing phenotype. However, the genetic basis for this phenomenon is unclear. Therefore, effective methods are needed for selecting parents to create F1 hybrid varieties with the desired PS. In this study, we used bulked segregant analysis with F1 Ningyou 1179 (strong PS) and its F2 population, and through analyzing both parental haplotypes and PS data for 918 hybrid rice varieties, to identify the genetic basis of transgressive late maturation which is dependent on dominance complementation effects of Hd1, Ghd7, DTH8, and PRR37 from both parents rather than from a single parental genotype. We designed a molecular marker-assisted selection system to identify the genotypes of Hd1, Ghd7, DTH8, and PRR37 in parental lines to predict PS in F1 plants prior to crossing. Furthermore, we used CRISPR/Cas9 technique to knock out Hd1 in Ning A (sterile line) and Ning B (maintainer line) and obtained an hd1-NY material with weak PS while retaining the elite agronomic traits of NY. Our findings clarified the genetic basis of transgressive late maturation in hybrid rice and developed effective methods for parental selection and gene editing to facilitate the breeding of hybrid varieties with the desired PS for improving their adaptability.

Enhanced antioxidant and antibacterial activities of chitosan/zein nanoparticle Pickering emulsion-incorporated chitosan coatings in the presence of cinnamaldehyde and tea polyphenol.

Pickering emulsions were prepared by using zein/chitosan nanoparticles as stabilizer and then incorporated into chitosan coatings. To improve the stability and performances, tea polyphenol and cinnamaldehyde (CA) were used to modulate the formation and functionalities of Pickering emulsions. The oil phase in Pickering emulsions were set at 5 % and 20 % to alter the hydrophobicity of chitosan coatings. Physical, structural, antioxidant and antibacterial activities of chitosan coatings with Pickering emulsions were characterized. Tea polyphenol significantly enhanced antioxidant capacity of chitosan coatings from 2.09 % to 57.61 % of DPPH value and from 2.63 % to 38.85 % of ABTS value. CA effectively increased the antibacterial activity of chitosan coatings against S. aureus and E. coli. Under 20 % oil content, the inhibition zones on S. aureus and E. coli increased from 3.03 ± 0.23 mm to 18.39 ± 1.22 mm and 7.66 ± 1.61 mm to 15.70 ± 1.75 mm, respectively. The preservative effect of chitosan coatings on fresh pork was further confirmed that the shelf-life of fresh pork could be extended by >4 days. These results suggested a great potential application of Pickering emulsion-incorporated chitosan coatings in the preservation of fresh pork.

A wireless battery-free eye modulation patch for high myopia therapy.

The proper axial length of the eye is crucial for achieving emmetropia. In this study, we present a wireless battery-free eye modulation patch designed to correct high myopia and prevent relapse. The patch consists of piezoelectric transducers, an electrochemical micro-actuator, a drug microneedle array, µ-LEDs, a flexible circuit, and biocompatible encapsulation. The system can be wirelessly powered and controlled using external ultrasound. The electrochemical micro-actuator plays a key role in precisely shortening the axial length by driving the posterior sclera inward. This ensures accurate scene imaging on the retina for myopia eye. The drug microneedle array delivers riboflavin to the posterior sclera, and µ-LEDs' blue light induces collagen cross-linking, reinforcing sclera strength. In vivo experiments demonstrate that the patch successfully reduces the rabbit eye's axial length by ~1217 µm and increases sclera strength by 387%. The system operates effectively within the body without the need for batteries. Here, we show that the patch offers a promising avenue for clinically treating high myopia.

Safety and efficacy of stoma site selection in CT-guided percutaneous gastrostomy: a retrospective analysis.

PURPOSE: To compare the safety and efficacy of CPG in the rectus abdominis and intercostal regions. MATERIALS AND METHODS: This retrospective study included 226 patients who underwent CPG at a single center, with the stoma placed in the rectus abdominis or intercostal region. Surgical outcomes and complications, such as pain and infection within 6 months postoperatively, were recorded. RESULTS: The surgical success rate was 100%, and the all-cause mortality rate within 1 month was 0%. An intercostal stoma was placed in 56 patients; a rectus abdominis stoma was placed in 170 patients. The duration of surgery was longer for intercostal stoma placement (37.66 ± 14.63 min) than for rectus abdominis stoma placement (30.26 ± 12.40 min) (P = 0.000). At 1 month postsurgery, the rate of stoma infection was greater in the intercostal group (32.1%) than in the rectus abdominis group (20.6%), but the difference was not significant (P = 0.077). No significant difference was observed in the infection rate between the two groups at 3 or 6 months postsurgery (P > 0.05). Intercostal stoma patients reported higher pain scores during the perioperative period and at 1 month postsurgery (P = 0.000), but pain scores were similar between the two groups at 3 and 6 months postsurgery. The perioperative complication rates for intercostal and rectus abdominis surgery were 1.8% and 5.3%, respectively (P = 0.464), with no significant difference in the incidence of tube dislodgement (P = 0.514). Patient weight improved significantly at 3 and 6 months postoperatively compared to preoperatively (P < 0.05). CONCLUSION: Rectus abdominis and intercostal stomas have similar safety and efficacy. However, intercostal stomas may result in greater short-term patient discomfort.

Structural characteristics and influencing factors of a spatial correlation network for tourism environmental efficiency in China.

The tourism industry in China presents uneven tourism efficiency but deepening spatial associations; thus, tourism resources must be more rationally allocated. In this study, the highly efficient SBM model was used to measure the tourism environmental efficiency of 31 provinces in China. A spatial correlation network is then constructed based on the gravity model, and the structural characteristics and influencing factors of the network are analyzed. The results show that (1) the overall tourism environmental efficiency in China presents a fluctuating growth trend, with significantly higher values observed in the eastern region than in the central and western regions; moreover, the growth in efficiency in the eastern region has been relatively stable in recent years, that in the central region has increased, while that in the western region has significantly declined. (2) A spatially linked network with a stable tourism environmental efficiency structure has been formed in China. The number of network relations and density of the network fluctuate and increase, while the network efficiency continues to decrease; however, a strong small-world nature is observed. (3) An obvious network core-edge structure is observed, with Shanghai, Beijing, Zhejiang, and Jiangsu at the center showing a significant intermediary role and remote provinces such as Tibet, Xinjiang, Ningxia, and Inner Mongolia at the edge showing fewer connections. (4) The four major plates of China based on the CONCOR algorithm are sparsely connected internally and show strong inter-plate connections and spillover effects. (5) The industry support capacity difference matrix, technological development level difference matrix, transportation accessibility difference matrix, and environmental regulation level difference matrix significantly and positively affect spatial associations, while the geographical distance matrix significantly and negatively affect spatial association relationship establishment. These findings have important theoretical and practical significance for the sustainable development of tourism in China's provinces and cities.

Spatiotemporal differences and influencing factors of urban vitality and urban expansion coupling coordination in the Pearl River Delta.

In the contemporary state of accelerated urbanization in China, urban expansion has become the mainstream trend of urban development. However, in the context of urban expansion, whether coordinated development between urban expansion and urban vitality is achieved is an important concern facing the current urban expansion process. The Pearl River Delta (PRD) urban agglomeration is a hotspot of rapid economic development in China, and it is essential to examine the coupling relationship between the region's urban dynamics and urban expansion to promote rational urban development. This study analyzes the dynamics and evolutionary characteristics of urban expansion based on multisource nighttime light data, employing an UERI and other methods. We apply the entropy method and coupling coordination model to evaluate urban vitality and conduct a coupling analysis of urban expansion and urban vitality from 1992 to 2021. The results show that during the study period, the urban built-up area of the PRD urban agglomeration increased by 61,131.59 km2, urban vitality gradually increased each year, the coupling coordination between urban vitality and urban expansion gradually increased, and factors such as economic development, urban planning, and geographical location advantages also influenced changes in urban vitality and urban expansion and their coupling and coordination in the study area. This study provides a methodological reference and data support for the investigating the spatiotemporal evolution of urban agglomerations and urban vitality analysis.

Wave absorbing properties of Ni Nanoparticle/CNT composite film fabricated by AAO/CNTs electrode.

An effective wave absorbing Nano-Ni/carbon nanotubes (CNTs) composite film was developed by electrodeposition using an anodic aluminum oxide (AAO)/CNTs electrode. Scanning electron microscopy images confirmed the uniform dispersion of Ni nano-particles within the CNTs, and the particle diameter increasing from 20 nm to 100 nm as the deposition time increased. XRD test results revealed that the crystal phase of the Ni nano-particles remained unchanged during different deposition time, exhibiting a Face Center Cubic (fcc) structure. The microwave electromagnetic properties of the film were evaluated using a vector network analyzer, and the return loss curve demonstrated that the Ni nano-particles/CNTs composite exhibited exceptional wave absorption capabilities. The composite film showed an effective absorption width of 13 GHz (4-17 GHz) and achieved a minimum reflection loss (RL) of -17 dB at 14 GHz.

Texture Analysis of Chinese Dried Noodles during Drying Based on Acoustic-Mechanical Detection Methods.

To better understand the textural transformation of Chinese dried noodles during the drying process, a convenient acoustic-force detection method was established. By comparing the breaking point, it was possible to determine the time-scale correlation between the force-displacement curves and acoustic spectrograms. The acoustic eigenvalues showed a consistent upward trend with the mechanical parameters during the drying process. With a wave crest reaching 152.8 dB and a signal maximum reaching 0.072, the structural stability of the dried noodles hardly induces a higher acoustic response. This suggests that the mechanical strength and rigidity of the dried noodles undergo minimal changes during this period. In comparison to the mechanical parameters, the acoustic eigenvalues accurately describe the changes in texture of dried noodles under various drying conditions, moreover, the sound threshold also provides a more effective response to the dried noodles' structural strength threshold. Therefore, the acoustic detection method can be applied to assist the conventional mechanical measurement in the field of the texture evaluation of dried food.

Effect of two different doses of nalbuphine for postoperative analgesia in children with cleft palate: a randomized controlled trial.

BACKGROUND: Cleft palate repair surgery may result in severe pain in the immediate postoperative period. The aim of this study is to compare the effects of different doses of nalbuphine for postoperative analgesia in children with cleft palate. METHODS: From November 2019 to June 2021, 90 children (45 males and 45 females, age 9-20 months old, ASA class I-II) were selected for palatoplasty. They were randomly divided into three groups: the control group (Group C), the N1 group (postoperative analgesia with 0.05 mg/kg/h nalbuphine) and the N2 group (postoperative analgesia with 0.075 mg/kg/h nalbuphine). Each group had 30 cases. Nalbuphine was not continuously infused in Group C but was continuously infused in Groups N1 and N2 at rates of 0.05 mg/kg/h and 0.075 mg/kg/h, respectively, for 24 h for postoperative analgesia. The FLACC analgesia score and Ramsay Sedation score were recorded at 10 min (T1), 30 min (T2), 2 h (T3), 12 h (T4) and 24 h (T5) after the operation. Adverse reactions such as nausea, vomiting and respiratory depression were observed and recorded. RESULTS: Compared with those in Group C, the FLACC scores in the N1 and N2 groups decreased significantly at T1-T5 (p < 0.05); the Ramsay Sedation score in the N1 group was significantly higher at T3 and T4 (p < 0.05), and that in the N2 group was significantly higher at T1-T5 (p < 0.05). Compared with that in the N1 group, the FLACC score in the N2 group was not significantly different, and the Ramsay Sedation score increased significantly at T5 (p < 0.05). CONCLUSION: Using 0.05 mg/kg/h Nalbuphine continuously for 24 h for postoperative analgesia in children with cleft palate has a better effect and fewer adverse reactions. TRIAL REGISTRATION: This study was registered at ChiCTR1900027385 (11/11/2019).

Early cumulus cell removal increases cumulative live birth rate while having no negative effect on the malformation rate in in vitro fertilization: a propensity score-matched cohort study.

PURPOSE: The aim of this study was to investigate the efficacy and safety of early cumulus cell removal (ECCR) during human in vitro fertilization (IVF). METHODS: A retrospective analysis was performed between January 2011 and December 2019. The study enrolled 1131 couples who underwent IVF treatment with ECCR. After propensity score matching at a 1:1 ratio, 1131 couples who underwent overnight coincubation of gametes were selected. The main outcome measure was the cumulative live birth rate. Secondary outcome measures included the cumulative pregnancy rate, polyspermy rate, available embryo rate, miscarriage rate, malformation rate, time to live birth, and oocyte-to-baby rate. RESULTS: There were no significant differences found between the two groups in the polyspermy rate, available embryo rate, miscarriage rate, time to live birth, oocyte-to-baby rate, and neonatal congenital anomalies rate. The results of the study showed that ECCR was associated with a significantly higher cumulative live birth rate and cumulative pregnancy rate, along with a significantly lower fertilization rate. CONCLUSIONS: ECCR tended to confer increased cumulative live birth rate and had no negative effect on the neonatal malformation rate.

Applicability of the FDA-approved Immunohistochemical Panel for Identification of MMRd Phenotype in Uterine Endometrioid Carcinoma.

Recently, the US Food and Drug Administration (FDA) approved the Ventana MMR RxDx Panel as the first immunohistochemical companion diagnostic test for identification of tumors with mismatch repair (MMR) status. The aim of this study was to investigate the accuracy of this test in comparison with polymerase chain reaction (PCR)-based microsatellite instability (MSI) analysis. We assessed the MMR/MSI concordance rate in 140 cases of endometrioid carcinoma. MMR status was evaluated by immunohistochemistry (MMR-IHC), and MSI status was evaluated by PCR-based analysis (MSI-PCR). Potential molecular mechanisms responsible for MSH6 staining variations were also analyzed. Immunohistochemistry showed that 34 tumors (24.3%) were MMRd; these included 26 with combined MLH1/PMS2 loss, 2 with combined MSH2/MSH6 loss, and 6 with isolated MSH6 loss. Heterogeneous MSH6 loss was found in 10 tumors and was recognized only in tumors with combined MLH1/PMS2 loss. Eight of 10 tumors with heterogeneous MSH6 loss harbored MSH6 C8 tract instability, suggesting a secondary somatic event after MLH1/PMS2 loss. MSI-PCR revealed that 102 tumors were MSS, 4 were MSI-low, and 34 were MSI-high. Consequently, MMR-IHC and MSI-PCR showed perfect concordance (kappa=0.080, P <0.0001). However, 10 of the 34 MSI-high tumors, including the 6 tumors with isolated MSH6 loss, showed only minimal microsatellite shift by MSI-PCR, which may have been erroneously interpreted as MSS or MSI-low. On the basis of these findings, we consider that the FDA-approved immunohistochemical panel can detect MMR variations consistently and is more accurate than MSI-PCR for determining the applicability of immune checkpoint inhibitors for treatment of endometrioid carcinomas.

Predicting subsolid pulmonary nodules before percutaneous needle biopsy: a comparison of artificial neural network and biopsy results.

AIM: To establish an artificial neural network (ANN) model to predict subsolid nodules (SSNs) before percutaneous core-needle biopsy (PCNB). The results of the two methods were compared to provide guidance on the treatment of SSNs. MATERIALS AND METHODS: This was a single-centre retrospective study using data from 1,459 SSNs between 2013 and 2021. The ANN was developed using data from patients who underwent surgery following computed tomography (CT) (SFC) and validated using data from patients who underwent surgery following biopsy (SFB). The prediction results of the ANN for the PCNB group and the histopathological results obtained after biopsy were compared with the histopathological results of lung nodules in the same group after surgery. Additionally, the choice of predictors for PCNB was analysed using multivariate analysis. RESULTS: There was no significant difference between the accuracies of the ANN and PCNB in the SFB group (p=0.086). The sensitivity of PCNB was lower than that of the ANN (p=0.000), but the specificity was higher (p=0.001). PCNB had better diagnostic ability than the ANN. The incidence of precursor lesions and non-neoplastic lesions in the SFB group was lower than that in the SFC group (p=0.000). A history of malignant tumours, size (2-3 cm), volume (>400 cm3) and mean CT value (≥-450 HU) are important factors for selecting PCNB. CONCLUSIONS: Both ANN and PCNB have comparable accuracy in diagnosing SSNs; however, PCNB has a slightly higher diagnostic ability than ANN. Selecting appropriate patients for PCNB is important for maximising the benefit to SSN patients.

Improved STMask R-CNN-based defect detection model for automatic visual inspection of an optics lens.

A lens defect is a common quality issue that has seriously harmed the scattering characteristics and performance of optical elements, reducing the quality consistency of the finished products. Furthermore, the energy hotspots coming from the high-energy laser through diffraction of optical component defects are amplified step by step in multi-level laser conduction, causing serious damage to the optical system. Traditional manual detection mainly relies on experienced workers under a special light source environment with high labor intensity, low efficiency, and accuracy. The common machine vision techniques are incapable of detecting low contrast and complex morphological defects. To address these challenges, a deep learning-based method, named STMask R-CNN, is proposed to detect defects on the surface and inside of a lens in complex environments. A Swin Transformer, which focuses on improving the modeling and representation capability of the features in order to improve the detection performance, is incorporated into the Mask R-CNN in this case. A challenge dataset containing more than 3800 images (18000 defect sample targets) with five different types of optical lens defects was created to verify the proposed approach. According to our experiments, the presented STMask R-CNN reached a precision value of 98.2%, recall value of 97.7%, F1 score of 97.9%, mAP@0.5 value of 98.1%, and FPS value of 24 f/s, which outperformed the SSD, Faster R-CNN, and YOLOv5. The experimental results demonstrated that the proposed STMask R-CNN outperformed other popular methods for multiscale targets, low contrast target detection and nesting, stacking, and intersecting defects sample detection, exhibiting good generalizability and robustness, as well as detection speed to meet mechanical equipment production efficiency requirements. In general, this research offers a favorable deep learning-based method for real-time automatic detection of optical lens defects.