FOOD CHEMISTRY
·
2025
·
IF 9.8
Fujia Dong,Zhaoyang Ma,Ying Xu,Yingjie Feng,Yingkun Shi,Hui Li,Fukang Xing,Guangxian Wang,Zhongxiong Zhang,Weiguo Yi,Songlei Wang
Veterinary drug residues in meat seriously harm human health. Rapid and accurate detection of veterinary drug residues is necessary to minimize contamination. Taking ofloxacin (OFX) residues in mutton as an example, the near-infrared hyperspectral imaging combined with explainable AI was used to evaluate the importance of feature wavelengths in the convolutional neural network-stacked sparse auto-encoder (CNN-SSAE) model for chemical properties. Based on this, the qualitative (residue identification-residue level identification) and quantitative detection of OFX residues in mutton was realized. The results showed that the accuracy of CNN-SSAE in identifying residue and residue level of OFX was 100% and 93.65%, respectively, and the correlation coefficients for validation (R 2 P ) in quantitative detection of OFX residue was 0.8980. In addition, SHapley Additive exPlanation (SHAP) values were used to identify feature wavelengths that contribute the most in the CNN-SSAE model, which effectively explained the quality attribute information that spectral and chemical values may improve the predicted results in the model decision process. The reliability of the CNN-SSAE model was evaluated by statistical validation methods (F-test and T-test). Finally, the visualization diagram of OFX content distribution was established. This study provides a method reference for explainability detection of veterinary drug residues.
研究领域:
食品科学
·
人工智能
·
生物医学
✓ 已验证
产品关联
DOI
PubMed
高影响因子
ADVANCED MATERIALS
·
2025
·
IF 26.8
Yuyang Sun,Tao Chen,Dongsheng Li,Hanyang Li,Tianci Ji,Fengxia Wang,Lining Sun,Chengkuo Lee,Huicong Liu
Colonoscopic continuum robots often lack sensing capabilities, risking tissue damage. An ideal robot electronic skin should offer full-body coverage, multiplexing, stretchability, and multifunctionality, but integration is challenging due to the robot's elongated structure. This work presents a stretchable electronic armor (E-armor) with a 3D crosslinked structure that enables 300 mm full coverage while accomplishing multiplexed simultaneous tactile and strain sensing through bioinspired artificial synapse mechanisms. The E-armor integrates 48 tactile sensing points through bilayer co-electrode strategy, reducing wiring while combining triboelectric encoding intelligence with innovative stretchable triboelectric interlinked films (TIFs) to form a triboelectric artificial synapse that generates digitally encoded signal pairs upon contact. A convolutional neural network and long short-term memory network (CNN-LSTM) deep learning framework achieve 99.31% accuracy in identifying multi-point tactile signals. A sodium alginate/polyacrylamide/sodium chloride (SA/PAM/NaCl) conductive hydrogel serves as a strain sensing element, providing excellent stretchability and biocompatibility, and allowing precise inference of bending angles at 12 strain sensing edges. A compliance control strategy coordinates tactile and strain signals to autonomously adjust continuum robot postures while ensuring smooth operation. The digital twin-based 3D visualization interface enhances human–robot interaction by digitally reconstructing both tactile and strain feedback, enabling real-time visualization of the continuum robot's intracolonic posture.
研究领域:
生物医学工程
·
机器人学
·
材料科学
✓ 已验证
产品关联
DOI
PubMed
高影响因子
AICHE JOURNAL
·
2020
·Vol. 66
·期 8
·e16260
·
IF 3.519
Chong Shen,Qibo Zheng,Minjing Shang,Li Zha,Yuanhai Su
In this work, an automatic liquid–liquidtwo-phase flow pattern recognition platform was developed to help circumvent the difficulties in labor-intensive hydrodynamics studies. Trained by about 30,000 of human-labeled flow pattern images, a convolutional neural network was built with the expert-level ability in the flow pattern recognition tasks and then coupled with automatic pump feeding system and online high-speed camera monitoring system to realize the high-throughput experimentation platform for microchannels. Effects of important factors such as flow rate, viscosity, interfacial tension, and so on were studied, and different flow pattern maps were obtained. With these thousands of flow pattern data in hand, we eventually drew the generalized liquid–liquidtwo-phase flow map and then proposed the relatively prudent criteria for slug flow operation window in the microchannel. This study extended the applications of artificial intelligence on microreactor technology or microfluidics, and in particular facilitated understanding complex hydrodynamics and flow patterns.
研究领域:
微流体技术
·
流体力学
·
人工智能应用
✓ 已验证
产品关联
DOI
高影响因子
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
·
2021
·
IF 4.098
Jiaqi Hu,Yanqiu Zou,Biao Sun,Xinyao Yu,Ziyang Shang,Jie Huang,Shangzhong Jin,Pei Liang
Pesticide detection is of tremendous importance in agriculture, and Raman spectroscopy/Surface-Enhanced Raman Scattering (SERS) has proven extremely effective as a stand-alone method to detect pesticide residues. Machine learning may be able to automate such detection, but conventional algorithms require a complete database of Raman spectra, which is not feasible. To bypass this problem, the present study describes a transfer learning method that improves the algorithm's accuracy and speed to extract features and classify Raman spectra. The transfer learning model described here was developed through the following steps: (1) the classification model was pre-trained using an open-source Raman spectroscopy database; (2) the feature extraction layer was saved after training; and (3) the training model for the Raman spectroscopy database was re-established while using self-tested pesticides and keeping the feature extraction layer unchanged. Three models were evaluated with or without transfer learning: CNN-1D, Resnet-1D, and Inception-1D, and they have improved the accuracy of spectrum classification by 6%, 2%, and 3%, with reduced training time and increased curve smoothness. These results suggest that transfer learning can improve the feature extraction capability and therefore accuracy of Raman spectroscopy models, expanding the range of Raman-based applications where transfer learning model can be used to identify the spectra of different substances.
研究领域:
分析化学
·
机器学习
·
农业科学
✓ 已验证
DOI
PubMed
高影响因子
ANALYST
·
2021
·Vol. 146
·期 8
·2490-2498
·
IF 4.616
Mingrui Jiang,Ting Shu,Chao Ye,Jing Ren,Shengjie Ling
As with other proteins, the conformation of the silk protein is critical for determining the mechanical, optical and biological performance of materials. However, an efficient, accurate and time-efficient method for evaluating the protein conformation from Fourier transform infrared (FTIR) spectra is still desired. A set of convolutional neural network (CNN)-based deep learning models was developed in this study to identify the silk proteins and evaluate their relative content of each conformation from FTIR spectra. Compared with the conventional deconvolution algorithm, our CNN models are highly accurate and time-efficient, showing promise in processing massive FTIR data sets, such as data from FTIR imaging, and in quick analysis feedback, such as on-line and time-resolved FTIR measurements. We compiled an open-source and user-friendly graphical Python program that allows users to analyze their own FTIR data set, which can be from the silk protein or other proteins, for the encouragement and convenience of interested researchers to use the CNN models.
研究领域:
生物信息学
·
材料科学
·
光谱学
✓ 已验证
DOI
PubMed
高影响因子
Journal of Applied Spectroscopy
·
2025
·
IF 1
Yang Jianlian,Feng Weiwei,Cai Zongqi,Wang Huanqing,Liang Xinghui
Fluorescent dissolved organic matter (FDOM) — particularly tryptophan (Trp), tyrosine (Tyr), and humic acid (HA) — serves as a crucial indicator in environmental monitoring. This study introduced a novel quantitative analysis approach for analyzing three-dimensional excitation–emission matrix spectra (3D-EEMs) of FDOM using convolutional neural networks (CNNs). The performance of the CNN model was evaluated and compared with the self-weighting alternating trilinear decomposition (SWATLD) algorithm. Results revealed that the proposed model significantly outperforms the SWATLD algorithm. Specifically, the CNN model achieved R 2 , RMSE, and MAPE values of 0.964, 0.047, and 14.950%, respectively, while for the SWATLD algorithm, these values were 0.944, 0.062, and 17.439%. Augmentation of the original spectral dataset did not yield a substantial improvement in the performance of the SWATLD algorithm, but it significantly enhanced the prediction ability of the CNN model. This enhancement was evident in the improved R 2 , RMSE, and MAPE values of 0.989, 0.030, and 12.837%, highlighting the critical role of data augmentation in boosting the performance of the CNN model, especially when dealing with a limited dataset. Application of the CNN model to water samples from Laizhou Bay yielded satisfactory results, enabling a simple and rapid analysis of FDOM in seawater. Therefore, an accurate and convenient analytical model was developed based on EEMs and CNNs, which can swiftly determine the concentration of FDOM in the environment and provide valuable references for environmental monitoring and early warning.
研究领域:
环境科学
·
分析化学
·
人工智能
✓ 已验证
DOI
高影响因子
FOOD CONTROL
·
2025
·
IF 6.3
Yaqi Hu,Wei Sheng,Selorm Yao-Say Solomon Adade,Jun Wang,Huanhuan Li,Quansheng Chen
Tea polyphenols (TPs) and dihydromyricetin (DMY) are critical quality attributes of vine tea. This study developed a smartphone-based portable near infrared (NIR) device integrated with machine learning (ML) and deep learning (DL) approaches for rapid prediction of TPs and DMY in vine tea. NIR spectra of the vine tea samples were acquired using the developed portable device and smartphone software, while the contents of TPs and DMY were determined using UV–Vis spectrophotometer and high-performance liquid chromatography (HPLC). To accurately analyze the spectral data, various ML and DL models were evaluated and compared. Results indicate that DL models, including convolutional neural networks (CNN), long-short-term memory (LSTM) and CNN-LSTM, demonstrated superior predictive performance compared to traditional ML approaches in large sample environments. Thereinto, CNN-LSTM exhibited the optimal predictive performance for TPs (Rp = 0.9816; RPD = 5.24) and DMY (Rp = 0.9900; RPD = 7.11). Additionally, the optimal model 's validation performance are commendable, with a maximum coefficient of determination (0.9483 for TPs and 0.9625 for DMY). This demonstrates that the developed intelligent portable NIR instrument coupled with DL tools enables rapid on-site detection of vine tea quality components. Furthermore, it provides a potential strategy for real-time quality monitoring during vine tea online processing.
研究领域:
食品科学
·
机器学习
·
深度学习
✓ 已验证
产品关联
DOI
高影响因子
Nano Energy
·
2024
·
IF 16.8
Xinrong Zhi,Shifan Ma,Yifan Xia,Biao Yang,Siyu Zhang,Kangting Liu,Mingyuan Li,Shuhan Li,Wan Peiyuan,Xin Wang
Tactile sensor as core component of robotic haptics enables intelligent perception of external environment for smart robot. However, it is a critical challenge to realize simultaneously high-sensitivity pressure sensing and high-resolution tactile imaging of tactile sensor. Herein, a hierarchical hybrid tactile sensor array (HTSA), which is fabricated by vertically integrating top arrayed TENG layer and bottom arrayed capacitive pressure sensor layer, is developed to achieve simultaneous pressure sensing and tactile imaging for tactile pattern recognition. The capacitive pressure sensor array (8 × 8) can realize pressure sensing with a sensitivity of 238.73 kPa −1 . Meanwhile, the TENG-based tactile sensing array enables high-resolution tactile imaging of handwritten Chinese characters. More importantly, the HTSA-based tactile pattern recognition system (H-TPRS) is developed to achieve accurate user signature authentication. Based on handwriting behavior characteristics, an accuracy of 97.3% of tactile pattern recognition is achieved in identifying the different-user handwriting identical letters under assistance of convolutional neural network . Consequently, the H-TPRS can achieve high-sensitivity pressure sensing and high-resolution tactile imaging of Chinese characters for user authentication, which demonstrates potential application in the field of intelligent tactile perception .
研究领域:
机器人学
·
材料科学
·
人工智能
✓ 已验证
产品关联
DOI
高影响因子
Applied Sciences-Basel
·
2022
·Vol. 12
·期 12
·5843
·
IF 2.838
Xiangan Long,Chongjun Ma,Han Sheng,Liwen Chen,Yiyan Fei,Lan Mi,Dongmei Han,Jiong Ma
Subvisible particles are an ongoing problem in biotherapeutic injectable pharmaceutical formulations, and their identification is an important prerequisite for tracing them back to their source and optimizing the process. Flow imaging microscopy (FIM) is a favored imaging technique, mainly because of its ability to achieve rapid batch imaging of subvisible particles in solution with excellent imaging quality. This study used VGG16 after transfer learning to identify subvisible particle images acquired using FlowCam. We manually prepared standards for seven classes of particles, acquired the image information through FlowCam, and fed the images over 5 µm into VGG16 consisting of a convolutional base of VGG16 pre-trained with ImageNet data and a custom classifier for training. An accuracy of 97.51% was obtained for the test set data. The study also demonstrated that the recognition method using transfer learning outperforms machine learning methods based on morphological parameters in terms of accuracy, and has a significant training speed advantage over scratch-trained CNN. The combination of transfer learning and FIM images is expected to provide a general and accurate data-analysis method for identifying subvisible particles.Keywords:image analysis;protein formulation;convolutional neural network;particle identification;machine learning
研究领域:
药学
·
生物信息学
✓ 已验证
产品关联
DOI
高影响因子
MICROCHEMICAL JOURNAL
·
2023
·
IF 4.8
Kaiqiang Wang,Zilin Yue,Hong Lin,Qi Wang,Lei Wang,Ye Tian,Lihui Ren
Freshness is a key indicator for assessing the nutritional and safety qualities of fish products. This study aims to apply the convolutional neural network (CNN) for modeling the Raman spectra data, for the rapid classification of the freshness grades of sea bass ( Lateolabrax japonicus ) fillets. Results suggested that the classification accuracy of CNN model was superior to principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), and support vector machine (SVM) based on Raman spectra data. In addition, analysis of variance (ANOVA) was demonstrated effective for extracting the Raman spectral features for CNN modeling. Based on the feature-selected CNN model using ANOVA, the average accuracy for identifying sea bass fillets freshness was improved to 90.6% and the testing time was reduced to 0.44 s. In particular, the classification accuracy for the second-grade freshness was significantly improved to 80.2% compared with the CNN model with full-band Raman spectra. This research provided an exemplar for combining the portable Raman spectrometer with deep learning for the rapid and non-destructive analysis of food quality.
研究领域:
食品科学
·
机器学习
·
光谱学
✓ 已验证
产品关联
DOI
高影响因子
Journal of Materials Chemistry C
·
2023
·Vol. 11
·期 43
·15233-15245
·
IF 6.4
Wei Xu,Junqi Cui,Fengze Bai,Longjiang Zheng,Chunhai Hu,Zhiguo Zhang,Zhen Sun,Yungang Zhang
The accuracy of luminescence thermometry is seriously hindered by thermal-induced luminescence quenching as well as traditional single spectral parameter-based analytical methods, which rely on subjective experience of humans and fail to effectively utilize the spectral features. Herein, thermally intensified luminescence of Cr3+ is successfully achieved in Gd3Ga5O12:Yb3+–Er3+–Cr3+ under 980 nm laser excitation and about 10-fold enhancement is observed at 853 K compared with that at 303 K. The reabsorption of Er3+ luminescence by Cr3+ and the phonon-assisted energy transfer from Er3+ to Cr3+ are responsible for the Cr3+ luminescence, and the latter is the key to the enhanced luminescence, which guarantees good signal-to-noise ratio of emissions at high temperatures. A convolutional neural network (CNN) is subsequently proposed to extract thermal information from the UC (upconversion) emissions, and the maximum error is just about 0.63 K in the temperature range of 303–853 K, along with an average error of only 0.15 K, much better than those obtained with conventional ratiometric approaches. Additionally, luminescence thermometry driven by CNNs can effectively resist the interference of background light and ensure measurement accuracy, further demonstrating the excellent robustness of the proposed thermometry strategy.
研究领域:
材料科学
·
光学
·
人工智能
✓ 已验证
产品关联
DOI
高影响因子
Applied Sciences-Basel
·
2024
·Vol. 14
·期 5
·2118
·
IF 2.5
Mengping Huang,Shuai Ma,Jinrong He,Wei Xue,Xueyan Hou,Yuqi Zhang,Xiaofeng Liu,Heping Bai,Ran Li
Amino acids found in minor coarse cereals are essential for human growth and development and play a crucial role in efficient and rapid quantitative detection. Surface-enhanced Raman spectroscopy (SERS) enables nondestructive, efficient, and rapid sample detection. Traditional SERS detection efficiency is constrained by the use of a single target. In this study, three different amino acids (cysteine, valine, and tryptophan) were detected simultaneously using a ZIF-8@AuNPs composite substrate. The linear range of detection was 10−3to 10−1M, with limits of detection (LODs) of 2.40 × 10−4M, 2.24 × 10−4M, and 1.55 × 10−4M, respectively. Same linear ranges and LODs were achieved with a one-dimensional convolutional neural network method. Furthermore, this substrate enabled the effective detection of amino acids in millet and efficient detection of cysteine in health products. This study presents a novel method for simultaneous detection of multiple analytes.
研究领域:
分析化学
·
生物医学工程
·
食品科学
✓ 已验证
产品关联
DOI
高影响因子
危险品化学品经营许可证（带存储）