RESUMO
The correlation between magnetic Barkhausen noise (MBN) features and the surface hardness of two types of die steels (Cr12MoV steel and S136 steel in Chinese standards) was investigated in this study. Back-propagation neural network (BP-NN) models were established with MBN magnetic features extracted by different methods as the input nodes to realize the quantitative prediction of surface hardness. The accuracy of the BP-NN model largely depended on the quality of the input features. In the extraction process of magnetic features, simplifying parameter settings and reducing manual intervention could significantly improve the stability of magnetic features. In this study, we proposed a method similar to the magnetic Barkhausen noise hysteresis loop (MBNHL) and extracted features. Compared with traditional MBN feature extraction methods, this method simplifies the steps of parameter setting in the feature extraction process and improves the stability of the features. Finally, a BP-NN model of surface hardness was established and compared with the traditional MBN feature extraction methods. The proposed MBNHL method achieved the advantages of simple parameter setting, less manual intervention, and stability of the extracted parameters at the cost of small accuracy reduction.
RESUMO
Experts usually inspect electroencephalogram (EEG) recordings page-by-page in order to identify epileptic seizures, which leads to heavy workloads and is time consuming. However, the efficient extraction and effective selection of informative EEG features is crucial in assisting clinicians to diagnose epilepsy accurately. In this paper, a determinant of covariance matrix (Cov-Det) model is suggested for reducing EEG dimensionality. First, EEG signals are segmented into intervals using a sliding window technique. Then, Cov-Det is applied to each interval. To construct a features vector, a set of statistical features are extracted from each interval. To eliminate redundant features, the Kolmogorov-Smirnov (KST) and Mann-Whitney U (MWUT) tests are integrated, the extracted features ranked based on KST and MWUT metrics, and arithmetic operators are adopted to construe the most pertinent classified features for each pair in the EEG signal group. The selected features are then fed into the proposed AdaBoost Back-Propagation neural network (AB_BP_NN) to effectively classify EEG signals into seizure and free seizure segments. Finally, the AB_BP_NN is compared with several classical machine learning techniques; the results demonstrate that the proposed mode of AB_BP_NN provides insignificant false positive rates, simpler design, and robustness in classifying epileptic signals. Two datasets, the Bern-Barcelona and Bonn datasets, are used for performance evaluation. The proposed technique achieved an average accuracy of 100% and 98.86%, respectively, for the Bern-Barcelona and Bonn datasets, which is considered a noteworthy improvement compared to the current state-of-the-art methods.
RESUMO
Gardenia blue is a natural blue pigment that is environmentally friendly, non-toxic, and stable. The hydrolysis of geniposide, catalyzed by ß-glucosidase, is a critical step in the production process of gardenia blue. However, ß-glucosidase is not resistant to high temperatures, limiting the production of gardenia blue. In this study, we investigated the effectiveness of a heat-resistant glucosidase obtained from Thermotoga maritima in the production of gardenia blue. The enzyme exhibited a maximum activity of 10.60 U/mL at 90 °C. Single-factor and orthogonal analyses showed that exogenously expressed heat-resistant glucosidase reacted with 470.3 µg/mL geniposide and 13.5 µg/mL glycine at 94.2 °C, producing a maximum yield of 26.2857 µg/mL of gardenia blue after 156.6 min. When applied to the dyeing of denim, gardenia blue produced by this method yielded excellent results; the best color-fastness was achieved when an iron ion mordant was used. This study revealed the feasibility and application potential of microbial production of gardenia blue.