Search

Chenpi, or sun-dried mandarin orange peel, held significant economic and medicinal value, yet counterfeit and substandard products were prevalent in the current market. The aging year of Chenpi was a crucial quality indicator, but accurately determining it through manual inspection was challenging. This study proposed a rapid, non-destructive method to discern the aging year of Chenpi by integrating hyperspectral imaging with deep learning. A total of 480 Chenpi samples across four aging years were collected, and their near-infrared hyperspectral data (wavelength range: 935.61~1720.23 nm) were obtained. A lightweight 1D-Rep network was utilized to develop a classification model enhanced by a feature band selection technique using multi-layer gradient-weighted class activation mapping (M-Grad-CAM). This approach evaluated the importance of spectral bands across multiple Rep-block layers, indicating band significance while considering inter-band and remote correlations. To validate the effectiveness of the proposed method, feature bands obtained from machine learning methods such as partial least squares discriminant analysis (PLS-DA), random forest (RF), and support vector machine (SVM) were used as comparative methods. The results showed that, the 1D-Rep full-spectrum model achieved an accuracy of 98.55%. When employing M-Grad-CAM for feature band selection and establishing a classification model based on the first nine feature bands, an accuracy greater than 90% could be achieved in feature band modeling. The accuracy reached 96.82% with 20 feature bands, significantly higher than that of the comparative models. This research effectively distinguishes Chenpi of different years using hyperspectral imaging technology, providing a methodological and theoretical basis for the development of portable detection instruments.

Sound event localization and detection have been applied in various fields. Due to the polyphony and noise interference, it becomes challenging to accurately predict the sound event and their occurrence locations. Aiming at this problem, we propose a Multiple Attention Fusion ResNet, which uses ResNet34 as the base network. Given the situation that the sound duration is not fixed, and there are multiple polyphonic and noise, we introduce the Gated Channel Transform to enhance the residual basic block. This enables the model to capture contextual information, evaluate channel weights, and reduce the interference caused by polyphony and noise. Furthermore, Split Attention is introduced to the model for capturing cross-channel information, which enhances the ability to distinguish the polyphony. Finally, Coordinate Attention is introduced to the model so that the model can focus on both the channel information and spatial location information of sound events. Experiments were conducted on two different datasets, TAU-NIGENS Spatial Sound Events 2020, and TAU-NIGENS Spatial Sound Events 2021. The results demonstrate that the proposed model significantly outperforms state-of-the-art methods under multiple polyphonic and noise-directional interference environments and it achieves competitive performance under a single polyphonic environment.

Conventional food fraud detection using hyperspectral imaging (HSI) relies on the discriminative power of machine learning. However, these approaches often assume a balanced class distribution in an ideal laboratory environment, which is impractical in real-world scenarios with diverse label distributions. This results in suboptimal performance when less frequent classes are overshadowed by the majority class during training. Thus, the critical research challenge emerges of how to develop an effective classifier on a small-scale imbalanced dataset without significant bias from the dominant class. In this paper, we propose a novel nondestructive detection approach, which we call the Dice Loss Improved Self-Supervised Learning-Based Prototypical Network (Proto-DS), designed to address this imbalanced learning challenge. The proposed amalgamation mitigates the label bias on the most frequent class, further improving robustness. We validate our proposed method on three collected hyperspectral food image datasets with varying degrees of data imbalance: Citri Reticulatae Pericarpium (Chenpi), Chinese herbs, and coffee beans. Comparisons with state-of-the-art imbalanced learning techniques, including the Synthetic Minority Oversampling Technique (SMOTE) and class-importance reweighting, reveal our method’s superiority. Notably, our experiments demonstrate that Proto-DS consistently outperforms conventional approaches, achieving the best average balanced accuracy of 88.18% across various training sample sizes, whereas the Logistic Model Tree (LMT), Multi-Layer Perceptron (MLP), and Convolutional Neural Network (CNN) approaches attain only 59.42%, 60.38%, and 66.34%, respectively. Overall, self-supervised learning is key to improving imbalanced learning performance and outperforms related approaches, while both prototypical networks and the Dice loss can further enhance classification performance. Intriguingly, self-supervised learning can provide complementary information to existing imbalanced learning approaches. Combining these approaches may serve as a potential solution for building effective models with limited training data.

Quality and feature quantity of low- and high-resolution image pairs for training are directly related to performance of learning-based super-resolution methods. In order to accurately increase learning features to the low- and high-resolution image pairs, we proposed a gradient degradation model of terahertz(THz) images to describe the process of a high-resolution THz image’s gradient map is how degraded to corresponding low-resolution THz image’s gradient map. And with the proposed model, we presented a low- and high-resolution THz image pair construction method with gradient fusion for learning-based super-resolution. As gradient maps are fused to training pairs, performance of learning-based super-resolution methods for THz images could be improved. In addition, we applied the low- and high-resolution THz image pair construction method with gradient fusion to very deep super-resolution(VDSR) method and zero shot super-resolution(ZSSR) method, named VDSR method with gradient fusion and ZSSR method with gradient fusion, respectively. To evaluate the performance of these two improved methods, comparison experiments with passive millimeter-wave images and our THz lab data are presented. The experimental results show that the performance of the VDSR method with gradient fusion and the ZSSR method with gradient fusion have significant improvements in peak signal-to-noise ratio and structural similarity index measure than the VDSR method and the ZSSR method, respectively. It indicates that performance of learning-based super-resolution methods for THz images could be improved by applying the proposed low- and high-resolution THz image pair construction method with gradient fusion.

Hyperspectral imaging combined with chemometric approaches is proven to be a powerful tool for the quality evaluation and control of fruits. In fruit defect-detection scenarios, developing an unsupervised anomaly detection framework is vital, as defect sample preparation is labor-intensive and time-consuming, especially for exploring potential defects. In this paper, a spectral–spatial, information-based, self-supervised anomaly detection (SSAD) approach is proposed. During training, an auxiliary classifier is proposed to identify the projection axes of principal component (PC) images that were transformed from the hyperspectral data cubes. In test time, the fully connected layer of the learned classifier was used as a ‘spectral–spatial’ feature extractor, and the feature similarity metric was adopted as the score function for the downstream anomaly evaluation task. The proposed network was evaluated with two fruit data sets: a strawberry data set with bruised, infected, chilling-injured, and contaminated test samples and a blueberry data set with bruised, infected, chilling-injured, and wrinkled samples as anomalies. The results show that the SSAD yielded the best anomaly detection performance (AUC = 0.923 on average) over the baseline methods, and the visualization results further confirmed its advantage in extracting effective ‘spectral–spatial’ latent representation. Moreover, the robustness of SSAD is verified with the data pollution experiment; it performed significantly better than the baselines when a portion of anomalous samples was involved in the training process.

Adulteration in dairy products has received world-wide attention, and at the same time, near infrared (NIR) spectroscopy has proven to be a promising tool for adulteration detection given its advantages of real-time response and non-destructive analysis. Regardless, the accurate and robust NIR model for adulteration detection is hard to achieve in practice. Convolutional neural network (CNN), as a promising deep learning architecture, is difficult to apply to such chemometrics tasks due to the high risk of overfitting, despite the breakthroughs it has made in other fields. In this paper, the ensemble learning method based on CNN estimators was developed to address the overfitting and random initialization problems of CNN and applied to the determination of two infant formula adulterants, namely hydrolyzed leather protein (HLP) and melamine. Moreover, a probabilistic wavelength selection method based on the attention mechanism was proposed for the purpose of finding the best trade-off between the accuracy and the diversity of the sub-models in ensemble learning. The overall results demonstrate that the proposed method yielded superiority regression performance over the comparison methods for both studied data sets, and determination coefficients (R2) of 0.961 and 0.995 were obtained for the HLP and the melamine data sets, respectively.

Near-infrared (NIR) spectroscopy and least squares support vector machine (LS-SVM) were applied to detect hydrolysed leather protein (HLP) adulteration in infant formula. In the classification approach, principal components (PCs) extracted by principal component analysis (PCA) were used as the inputs of the LS-SVM class model. The correct classification rates of authentic samples and adulterated samples in validation set were 96.7 % and 95.6 %, respectively. In the regression approach, PCs extracted by PCA were also utilized as inputs of LS-SVM to develop PCs-LS-SVM regression models. The optimal PCs-LS-SVM model was obtained with 4 PCs and the root mean squared error of calibration and prediction were 1.02 g·kg-1 and 1.46 g·kg-1, respectively. Moreover, sensitive wavelengths (SWs) were selected according to the regression coefficients obtained by partial least squares regression model and utilized as the input data of LS-SVM to develop SWs-LS-SVM regression models. The optimal SWs-LS-SVM model was obtained with 19 SWs and the root mean squared error of calibration and prediction were 0.64 g·kg-1 and 1.18 g·kg-1, respectively. The results demonstrated that NIR spectroscopy combined with LS-SVM could be applied as a fast, simple and nondestructive way to detect HLP adulteration in infant formula. [ABSTRACT FROM AUTHOR]

Hydrolyzed leather protein (HLP), which is collagen extracted from leather scraps, was intentionally added into dairy products by some illegal dairy producers and has caused food safety scares in China. In this study, near-infrared (NIR) spectroscopy and chemometric analysis were employed to detect HLP adulteration in infant formula. Infant formula adulterated with HLP powder can be well discriminated from authentic samples by using soft independent modelling of class analogy (SIMCA) method. Furthermore, partial least squares regression (PLS) and support vector regression (SVR) models were developed, optimized and compared for quantitative analyses of HLP adulteration concentrations. The SVR modelling method has demonstrated a better predicted capacity than PLS, and a relatively high (7.42) ratio of prediction (RPD) can be achieved. NIR spectroscopy coupled with the chemometric analysis technique could be a useful tool for simple, rapid, economical and non-destructive detection of HLP adulteration in dairy products. [ABSTRACT FROM AUTHOR]

Objective: Heart sound classification still suffers from the challenges involved in achieving high accuracy in the case of small samples. Dimension reduction attempts to extract low-dimensional features with more discriminability from high-dimensional spaces or raw data, and is popular in learning predictive models that target small sample problems. However, it can also be harmful to classification, because any reduction has the potential to lose information containing category attributes. Approach: For this, a novel SNMFNet classifier is designed to directly associate the dimension reduction process with the classification procedure used for promoting feature dimension reduction to follow the approach that is beneficial for classification, thus making the low-dimensional features more distinguishable and addressing the challenge facing heart sound classification in small samples. Main results: We evaluated our method and representative methods using a public heart sound dataset. The experimental results demonstrate that our method outperforms all comparative models with an obvious improvement in small samples. Furthermore, even if used with relatively sufficient samples, our method performs at least as well as the baseline that uses the same high-dimensional features. Significance: The proposed SNMFNet classifier significantly to improves the small sample problem in heart sound classification. [ABSTRACT FROM AUTHOR]