EResNet-SVM: an overfitting-relieved deep learning model for recognition of plant diseases and pests.

BACKGROUND: The accurate recognition and early warning for plant diseases and pests are a prerequisite of intelligent prevention and control for plant diseases and pests. As a result of the phenotype similarity of the hazarded plant after plant diseases and pests occur, as well as the interference of the external environment, traditional deep learning models often face the overfitting problem in phenotype recognition of plant diseases and pests, which leads to not only the slow convergence speed of the network, but also low recognition accuracy. RESULTS: Motivated by the above problems, the present study proposes a deep learning model EResNet-support vector machine (SVM) to alleviate the overfitting for the recognition and classification of plant diseases and pests. First, the feature extraction capability of the model is improved by increasing feature extraction layers in the convolutional neural network. Second, the order-reduced modules are embedded and a sparsely activated function is introduced to reduce model complexity and alleviate overfitting. Finally, a classifier fused by SVM and fully connected layers are introduced to transforms the original non-linear classification problem into a linear classification problem in high-dimensional space to further alleviate the overfitting and improve the recognition accuracy of plant diseases and pests. The ablation experiments further demonstrate that the fused structure can effectively alleviate the overfitting and improve the recognition accuracy. The experimental recognition results for typical plant diseases and pests show that the proposed EResNet-SVM model has 99.30% test accuracy for eight conditions (seven plant diseases and one normal), which is 5.90% higher than the original ResNet18. Compared with the classic AlexNet, GoogLeNet, Xception, SqueezeNet and DenseNet201 models, the accuracy of the EResNet-SVM model has improved by 5.10%, 7%, 8.10%, 6.20% and 1.90%, respectively. The testing accuracy of the EResNet-SVM model for 6 insect pests is 100%, which is 3.90% higher than that of the original ResNet18 model. CONCLUSION: This research provides not only useful references for alleviating the overfitting problem in deep learning, but also a theoretical and technical support for the intelligent detection and control of plant diseases and pests. © 2024 Society of Chemical Industry.

Evaluation of choroidal thickness and retinal nerve fiber layer thickness in Chinese pregnant women and healthy non-pregnant women.

Purpose: To evaluate choroidal thickness and retinal nerve fiber layer (RNFL) thickness in different trimesters using enhanced depth imaging optical coherence tomography (EDI-OCT). Methods: A prospective comparative study included 45 healthy pregnant women in the first trimester, 45 women in the second, 45 women in the third and 45 healthy non-pregnant women as the control group. Macular choroidal thickness was measured at three locations: The subfoveal, 1 â€‹mm temporal, and 1 â€‹mm nasal from the fovea with EDI-OCT. Peripapillary choroidal thickness (PPCT) and RNFL thickness parameters were automatically calculated by the Spectralis OCT. Results: The subfoveal, temporal and nasal macular choroidal thickness were all significantly thicker in the second trimester, compared with those parameters in the first, the third trimesters and the control group (all P â€‹< â€‹0.05). The PPCT was significantly increased in the second trimeter compared with the control group at global, temporal, temporal inferior, nasal and nasal inferior positions (all P â€‹< â€‹0.05). The RNFL thickness was also significantly increased in pregnant women at nasal superior and nasal inferior quadrants (all P â€‹< â€‹0.05). Conclusions: The choroidal thickness in pregnant women was found to be thicker than the control group, regardless of macular or optic disc location. Findings of RNFL thickening might indicate subclinical involvement of the central nervous system.

Semisynthesis, anti-oomycete and anti-fungal activities of ursolic acid ester derivatives.

Using ursolic acid (UA) as the lead compound, thirteen UA ester derivatives (3 and 7a-l) were synthesized by modifying their C-3 and C-28 positions, respectively, and their structures were well characterized by 1H NMR, 13C NMR, HRMS and melting points. Furthermore, we evaluated the anti-oomycete and anti-fungal activities of these compounds against Phytophthora capsici and Fusarium graminearum in vitro. The results showed that compound 7h exhibited prominent anti-oomycete and anti-fungal activities, and the median effective concentration (EC50) values of 7h against P. capsici and F. graminearum were 70.49 and 113.21 mg/L, respectively. This study suggested that the anti-oomycete and anti-fungal activities of esters synthesized by introducing acyloxy group at C-3 position of UA was more conspicuous than that of esters synthesized by introducing benzyloxy group at C-28 position. This result will pave the way for further modification of UA to develop potential new fungicides.

Sensitive and Enzyme-Free Pathogenic Bacteria Detection Through Self-Circulation of Molecular Beacon.

This research exhibits the design of a feasible, enzyme-free and sensitive fluorescent sensing assay for the detection of Staphylococcus aureus (S. aureus), using self-circulation of molecular beacons. With protein A on S. aureus as identifying target, the capture probe binds on the surface of S. aureus based on interaction between its aptamer section and protein A. Recognition of protein A by aptamer section in capture probe leads to allosterism of capture probe, exposing initiator section to activate the following self-circulation. After multiple circulation-based signal amplification, the method exhibits a favorable detection sensitivity and shows a promising prospect for the keratitis-related pathogenic bacteria detection. The highlights of the sensing assay are as follows: (i) capture probe is designed with aptamer section which endows the method a high selectivity; (ii) signal of bacteria is converted to nucleic acid signal after recognition of target bacteria by capture probe; and (iii) high sensitivity of method is derived from the self-circulation process. Therefore, we believe that the strategy can provide a useful platform for target bacteria detection and thus contribute to the diagnosis of infectious diseases.

Nanoparticle carriers based on copolymers of poly(ε-caprolactone) and hyperbranched polymers for drug delivery.

Novel amphiphilic copolymers based on poly(ε-caprolactone) (PCL) and hyperbranched poly (amine-ester) (HPAE) with various compositions were synthesized. The amphiphilic copolymers can self-assemble into nanoscopic micelles and their hydrophobic cores can encapsulate doxorubicin (DOX) in aqueous solutions. The DOX-loaded HPAE-co-PCL nanoparticles diameter increased from 121 to 184 nm with the increasing PCL segment in the copolymer composition. An in vitro study at 37°C demonstrated that DOX-release from nanoparticles at pH 5.0 was much faster than that at pH 7.4. The cytotoxicity for HeLa cells study demonstrated that DOX-loaded HPAE-co-PCL nanoparticles exhibited the anti-tumor effect was enhanced significantly, suggesting that the DOX-loaded HPAE-co-PCL nanoparticles have great potential as a tumor drug carrier.

Hydroxypropyl-ß-cyclodextrin copolymers and their nanoparticles as doxorubicin delivery system.

A novel biodegradable amphiphilic copolymer composed of hydroxypropyl-ß-cyclodextrin, polylactide, and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine, was successfully synthesized. The chemical structures of copolymers were determined by FT-IR, (1)H nuclear magnetic resonance (NMR) spectroscopy, (13)C NMR, (31)P NMR, thermogravimetric analysis ,and differential scanning calorimetry. Doxorubicin (DOX)-loaded copolymer nanoparticles (NPs) were prepared by double emulsion and nanoprecipitation methods. The factors of copolymer composition and fabrication methods, which influence size and encapsulation efficiency (EE) were investigated. Their EE to DOX could reach 90.6% at an available condition. In vitro release behavior of NPs showed a continuous release after a burst release. The antitumor activity of the DOX-loaded NPs against cancer HepG2 and A549 cells was evaluated by 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide method. The DOX-loaded copolymer NPs showed comparable anticancer efficacy with the free drug.

Amphiphilic hyper-branched co-polymer nanoparticles for the controlled delivery of anti-tumor agents.

In this investigation, we have designed and synthesized an amphiphilic co-polymer with hyper-branched poly(amine-ester) and polylactide (HPAE-co-PLA) to generate nanoparticles (NPs). These have been used to encapsulate a highly active hydrophobic anti-tumor agent, 2-benzoylpyridine 4-ethyl-3-thiosemicarbazone (Bp4eT). Encapsulation in NPs was done in an effort to increase the anti-tumor activity of this agent by facilitating its delivery to tumor cells. We have also examined and optimized the formulation parameters of the NPs that alter their drug-loading capacity and their physical, chemical and biological properties. The resulting NPs exhibited high Bp4eT-loading capacity and substantial stability in aqueous solution. In vitro drug release studies demonstrated a controlled drug release profile with increased release at acidic pH. Anti-tumor proliferation assays showed that both free drug and drug-encapsulated NPs markedly inhibited tumor cell proliferation in a time- and concentration-dependent manner. Direct microscopic observation revealed that the fluorescent NPs were taken up by cells and localized, in part, in organelles consistent with lysosomes. These results demonstrate a feasible application of the amphiphilic hyper-branched co-polymer, HPAE-co-PLA, as nanocarriers for intracellular delivery of potent anti-tumor agents.

[Preparation and drug release property of paclitaxel nanoparticles].

OBJECTIVE: To prepare paclitaxel nanoparticles and evaluate its drug release property in vitro. METHODS: The chitosan-poly (lactide) copolymers were synthesized by ring-opening polymerization. Paclitaxel nanoparticles were prepared by the nanoprecipitation method. The encapsulation efficiency of the nanoparticles and the in vitro release of paclitaxel were studied. RESULTS: The nanoparticles were fine and uniform spheres. The encapsulation efficiency of nanoparticles was up to 93.2%. The in vitro release profile showed the slow and stable release rate of paclitaxel. CONCLUSION: This method is simple and the nanoparticles possess good physical performance as sustained release character in vitro.

The shape evolution of gold seeds and gold@silver core-shell nanostructures.

We report the seed-dependent shape evolution of gold@silver (Au@Ag) core-shell nanostructures with various morphologies through using pre-existing Au nanocrystals as nuclei in a polyvinylpyrrolidone (PVP)-assisted polyol reduction process. Au nanocrystalline seeds with different shapes such as cube, truncated-octahedron, octahedron, twinned hexagon and triangle, five-twinned decahedron and nanorod are firstly synthesized by refluxing a 1,5-pentanediol solution containing Au precursors in the presence of PVP. The Au seeds obtained in this way then serve as the nuclei for further epitaxial growth of Ag shells by using Ag precursors via the same route. Scanning transmission electron microscope (STEM) characterization of the products obtained demonstrates that the morphological evolution of the Ag shells depends completely on the shapes of the Au seeds that are used. We have observed that the Au@Ag core-shell nanostructures formed with various regular shapes such as cube, bi-triangle, and nanorod with five-twinned cross section, are mostly surrounded by {100}-type Ag crystalline facets. Our findings provide new evidence and clear evolution routines from the Au cores with well-defined shapes to the corresponding Ag shells for the Au@Ag core-shell nanostructures by the family of the PVP-assisted polyol reduction methods.

Novel amphiphilic hyperbranched poly (amine-ester) copolymers nanoparticles as protein drug delivery.

Novel amphiphilic protein-loaded hyperbranched poly (amine-ester) copolymers nanoparticles were fabricated by double emulsion (DE) and nanoprecipitation (NP) methods. Encapsulation efficiency (EE) of nanoparticles to Bovine serum albumin (BSA) could reach 97.8% at BSA (0.01 g/ml), internal phase volume (0.5 ml), copolymer weight (50mg), and PVA concentration (4%, w/v) condition.