Background: Pumpkin seed oil is widely used to treat benign prostatic hyperplasia (BPH), a common disease in elder men. However, its active components and mechanism have remained to be elucidated. Objective: The objective of the present study was to investigate the active components of pumpkin seed oil and its mechanism against BPH. Design: Total phytosterol (TPS) was isolated from hull-less pumpkin (Cucurbita pepo L. var. Styriaca) seed oil and analyzed by gas chromatography/mass spectrometry (GC/MS). Three phytosterols were purified by preparative HPLC (high performance liquid chromatography) and confirmed by NMR (nuclear magnetic resonance). TPS (3.3 mg/kg body weight, 1 mL/day/rat) was administered intragastrically to the testosterone propionate-induced BPH rats for 4 weeks. The structure changes of prostate tissues were assessed by hematoxylin & eosin (H&E) staining. The expression of androgen receptor (AR) and steroid receptor coactivator 1 (SRC-1) was analyzed by immunohistochemistry, while that of 5α-reductase (5AR), apoptosis, or proliferation-related growth factors/proteins was detected by real-time quantitative polymerase chain reaction or western blotting. Results: The ∆7-phytosterols in TPS reached up to 87.64%. Among them, 24β-ethylcholesta-7,22,25-trienol, 24β-ethylcholesta-7,25(27)-dien-3-ol, and ∆7-avenasterol were confirmed by NMR. TPS treatment significantly ameliorated the pathological prostate enlargement and restored histopathological alterations of prostate in BPH rats. It effectively suppressed the expressions of 5AR, AR, and coactivator SRC-1. TPS inhibited the expression of proliferation-related growth factor epidermal growth factor, whereas it increased the expressions of apoptosis-related growth factor/gene transforming growth factor-β1. The proliferation-inhibiting effect was achieved by decreasing the ERK (extracellular signal-regulated kinase) phosphorylation, while apoptosis was induced by Caspase 3 activation through JNK (c-Jun N-terminal kinase) and p38 phosphorylation. Conclusion: TPS from hull-less pumpkin seed oil, with ∆7-phytosterols as its main ingredients, is a potential nutraceutical for BPH prevention.

The harvesting time of fresh tea leaves has a significant impact on product yield and quality. The aim of this study was to propose a method for real-time monitoring of the optimum harvesting time for picking fresh tea leaves based on machine vision. Firstly, the shapes of fresh tea leaves were distinguished from RGB images of the tea-tree canopy after graying with the improved B-G algorithm, filtering with a median filter algorithm, binary processing with the Otsu algorithm, and noise reduction and edge smoothing using open and close operations. Then the leaf characteristics, such as leaf area index, average length, and leaf identification index, were calculated. Based on these, the Bayesian discriminant principle and method were used to construct a discriminant model for fresh tea-leaf collection status. When this method was applied to a RGB tea-tree canopy image acquired at 45° shooting angle, the fresh tea-leaf recognition rate was 90.3%, and the accuracy for fresh tea-leaf harvesting status was 98% by cross validation. Hence, this method provides the basic conditions for future tea-plantation operation and management using information technology, automation, and intelligent systems.

No abstract

Deep neural networks are known to be vulnerable to adversarially perturbed inputs. A commonly used defense is adversarial training, whose performance is influenced by model capacity. While previous works have studied the impact of varying model width and depth on robustness, the impact of increasing capacity by using learnable parametric activation functions (PAFs) has not been studied. We study how using learnable PAFs can improve robustness in conjunction with adversarial training. We first ask the question: how should we incorporate parameters into activation functions to improve robustness? To address this, we analyze the direct impact of activation shape on robustness through PAFs and observe that activation shapes with positive outputs on negative inputs and with high finite curvature can increase robustness. We combine these properties to create a new PAF, which we call Parametric Shifted Sigmoidal Linear Unit (PSSiLU). We then combine PAFs (including PReLU, PSoftplus and PSSiLU) with adversarial training and analyze robust performance. We find that PAFs optimize towards activation shape properties found to directly affect robustness. Additionally, we find that while introducing only 1-2 learnable parameters into the network, smooth PAFs can significantly increase robustness over ReLU. For instance, when trained on CIFAR-10 with additional synthetic data, PSSiLU improves robust accuracy by 4.54% over ReLU on ResNet-18 and 2.69% over ReLU on WRN-28-10 in the ∞ threat model while adding only 2 additional parameters into the network architecture. The PSSiLU WRN-28-10 model achieves 61.96% AutoAttack accuracy, improving over the state-of-the-art robust accuracy on RobustBench . Overall, our work puts into context the importance of activation functions in adversarially trained models.

No abstract