SENSORS AND ACTUATORS B-CHEMICAL
·
2025
·
IF 7.7
Haixia Mei,Jingyi Peng,Tao Wang,Bowei Zhang,Fuzhen Xuan
Although current unsupervised domain adaptation (UDA) methods enable knowledge transfer without labeled data, their performance in regression tasks is often hindered by noisy pseudo-labels in the target domain. This challenge is particularly pronounced in gas concentration prediction, which demands higher label accuracy than classification tasks, yet lacks systematic research. To address this issue, this study proposes a novel semi-supervised cross-domain regression framework named Online-Teacher Domain-Adaptive Regression Network (OT-DARN) for accurate multi-gas concentration estimation. OT-DARN integrates an online teacher-student architecture that incorporates domain adversarial learning, along with a pseudo-label selection mechanism, to enhance both transferability and generalization. Extensive experiments on 15 commercially available gas sensor datasets demonstrate that OT-DARN achieves state-of-the-art performance with a minimum mean absolute error (MAE) of 0.7769 parts per million (ppm) under n-propanol concentrations of 10–100 ppm at (50 ± 10) % relative humidity and a temperature of (25 ± 2) °C, with a maximum R² of 0.9986 while reducing the labeling cost by 70 %. Even under fully unlabeled settings, the model achieves an MAE of 6.1206 and R² of 0.9317. Furthermore, bidirectional transfer experiments reveal no statistically significant differences between migration directions, validating its robustness. OT-DARN also consistently outperforms four representative UDA methods across various source–target combinations. Ablation studies further confirm the individual and joint contributions of key modules. This work pioneers semi-supervised cross-domain regression for gas concentration prediction and offers a cost-efficient, high-precision solution applicable to real-world sensing scenarios.
研究领域:
传感器与仪器
·
机器学习
·
环境监测
·
气体传感
·
迁移学习
·
回归分析
✓ 已验证
产品关联
DOI
高影响因子
ACS Applied Materials & Interfaces
·
2020
·
IF 8.758
Qiaofen Chen,Zhao Chen,Dong Liu,Zhengfu He,Jianmin Wu
A flexible electronic-nose (E-nose) was constructed by assembling graphene oxide (GO) using different types of metal ions (Mx+) with different ratio of GO to Mx+. Owing to the cross-linked networks, the Mx+-induced assembly of graphene films resulted in different porous structures. A chemi-resistive sensor array was constructed by coating the GO–M hybrid films on PET substrate patterned with 8 interdigited electrodes, followed by in situ reduction of GO to rGO with hydrazine vapor. Each of the sensing elements on the sensor array showed a cross-reactive response toward different types of gases at room temperature. Compared to bare rGO, incorporation of metal species into rGO significantly improved sensitivity owing to the additional interaction between metal species and gas analyte. Principle component analysis (PCA) showed that four types of exhaled breath (EB) biomarkers including acetone, isoprene, ammonia, and hydrothion in sub-ppm concentrations can be discriminated well. To overcome the interference from humidity in EB, a protocol to collect and analyze EB gases was established and further validated by simulated EB samples. Finally, clinical EB samples collected from patients with lung cancer and healthy controls were analyzed. In a 106 case study, the healthy group can be accurately distinguished from lung cancer patients by linear discrimination analysis. With the assistance of an artificial neural network, a sensitivity of 95.8% and specificity of 96.0% can be achieved in the diagnosis of lung cancer based on the E-nose. We also find that patients with renal failure can be distinguished through comparison of dynamic response curves between patient and healthy samples on some specific sensing elements. These results demonstrate the proposed E-nose will have great potential in noninvasive disease screening and personalized healthcare management.
研究领域:
医学诊断
·
材料科学
·
生物传感
✓ 已验证
DOI
高影响因子
SENSORS AND ACTUATORS B-CHEMICAL
·
2024
·
IF 8
Zengyi Tang,Yamin Liu,Mianyi Xiang,Daxiang Cui,Qichao Li
The detection of volatile organic compounds (VOCs) has numerous applications in environmental monitoring and biomedicine. Metal oxide semiconductor (MOS) gas sensors are widely used in gas detection due to their simple structure, low fabrication cost, and fast response speed. However, MOS gas sensors have broad-spectrum responsiveness to VOCs, making selectivity a challenge. The most commonly used method to obtain selectivity is the E-nose system consisting of multiple sensors. However, the size of E-noses has increased due to the number of sensors with high power consumption and complicated signal processing. In recent years, deep learning has flourished, allowing for accurate discrimination of VOCs and concentration prediction using E-nose consisting of a single sensor. This method significantly reduces power consumption and facilitates portable applications. In this work, we synthesize SnO 2 -ZnO composite oxide and fabricate a E-nose consisting of a single MEMS gas sensor. The response performance of the E-nose to VOCs is stable with a high output signal-noise ratio (SNR). We design a lightweight deep learning model based on SqueezeNet and train it using transfer learning. Linear regression is used to predict the concentration of VOCs. Our work achieves accurate identification and low-error concentration prediction of VOCs, which is valuable for subsequent portable VOC detection systems.
研究领域:
环境科学
·
生物医学工程
✓ 已验证
DOI
高影响因子
FOOD CHEMISTRY
·
2024
·
IF 8.5
Xiangxiang Sun,Yumei Yu,Zhenyu Wang,Kumayl Hassan Akhtar,Ahmed S.M. Saleh,Wenhao Li,Dequan Zhang
Effects of braising duration on volatile organic compounds (VOCs) and lipids in chicken were investigated. Aroma profiles identified by an electronic nose were effective in differentiating braising stages. During braising process, a total of 25 key VOCs were detected in braised chicken, and sample braised for 210 min exhibited the highest level of key VOCs. Additionally, a gas chromatography mass spectrometry fingerprint was established to evaluate the distribution of VOCs throughout the braising process. Partial least square discriminant analysis indicated that 2-heptanone, 3-methyl-2-butanone, octanal, nonanal, butanal, (E)-2-pentenal, 1-octen-3-ol, 1-hexanol, pentanal, hexanal, and 1-pentanol significantly affected flavor characteristics of braised chicken. Furthermore, 88 differential lipids were screened, and glycerolipids metabolic was found to be main metabolic pathway during braising process. Triglycerides (TG) and phosphatidyl ethanolamine (PE), such as TG (16:0/18:1/18:2), TG (18:0/18:1/18:2), TG (18:1/18:2/18:3), TG (18:1/18:1/18:2), PE (O-18:2/18:2), PE(O-18:2/18:1), and TG (16:0/16:1/18:2), played a vital role in the generation of VOCs.
研究领域:
食品科学
·
分析化学
·
代谢组学
✓ 已验证
DOI
PubMed
高影响因子
Frontiers in Chemistry
·
2026
·
IF 5.3
Pengfei Yang,Lingqi Kong,Qiongbo Wang,Qiang Liu,Xiujin Duan,Chen Hu,Zhengbo Feng,Qi Yang,Huabo Jv
Introduction To investigate the impact of different processing methods on the volatile components in Aglaia odorata flower extract (AOFEs). Methods Headspace solid-phase microextraction coupled with gas chromatography-mass spectrometry and electronic nose (E-Nose) analysis were employed to characterize volatiles of extracts obtained by ultrasound-assisted extraction (UAE), microwave-assisted extraction (MAE), and heated reflux extraction (HRE). Multidimensional assessment using aroma radar charts, orthogonal partial least squares-discriminant analysis (OPLS-DA), K-means clustering, and relative odor activity value (ROAV) revealed significant processing-dependent variations. Results and discussion The results indicated that the fiber coated with DVB/CAR/PDMS had optimal extraction efficiency. A total of 46 compounds were identified, including eight alcohols, four aldehydes, one acid, 25 terpenes, seven ketones, and one heterocyclic compound. UAE and MAE had 36 and 38 compounds respectively, sharing similar compositional profiles but differing in concentrations, while HRE produced only 25 compounds Sensory evaluation and E-Nose results revealed differences in the aroma profiles of the extracts, with UAE and MAE extracts exhibiting intensified floral and sweet notes, whereas HRE displayed prominent green and spicy characteristics. K-means clustering categorized volatile evolution trends into four distinct subclasses. OPLS-DA identified 13 differential volatiles with variable importance in projection greater than 1, with ROAV analysis further selecting eight key markers, including (1R,7 S )-germacra-4(15),5,10(14)-trien-1 β -ol, α -humulene, copaene, β -cadinene, (E)- β -caryophylene, Γ -cadinene, humulene oxide II, and caryophyllene oxide. These compounds collectively contribute to the sweet and floral attributes of AOFEs. This study elucidated extraction-method-dependent volatile profiles and aroma characteristics, providing theoretical guidance for process optimization and quality enhancement in AOFEs production.
✓ 已验证
产品关联
DOI
PubMed
高影响因子
Food Chemistry-X
·
2022
·
IF 6.443
LI Junxing,Miao Aiqing,ZHAO Gangjun,Liu Xiaoxi,Wu Haibin,Luo Jianning,Gong Hao,Zheng Xiaoming,Deng Liting,Ma Chengying
‘Taro-like’ aroma is a pleasant flavor and value-added trait in pumpkin species imparted by unknown key volatile compounds. In this study, we used the electronic nose (E-nose), gas chromatography-mass spectrometry (GC–MS), and GC-Olfactometry (GC-O) to study the aroma profile, volatile compounds, and key contributors, respectively. By E-nose and GC–MS, we found significant differences in the aroma profiles and volatile compounds between fruits from five samples with/without ‘taro-like’ aroma. According to the analysis of differential volatile compounds obtained from GC–MS and the GC-O analysis of the sample with ‘taro-like’ aroma, we found that 2-acetyl-1-pyrroline representing the ‘taro’ odor was only identified in the sample with ‘taro-like’ aroma. Therefore, we conclude that 2-acetyl-1-pyrroline is the key contributor to the 'taro-like' aroma. Moreover, the relationship between 2-acetyl-1-pyrroline and ‘taro-like’ aroma was further verified via other pumpkin samples. Our results provide a theoretical basis for understanding the aroma characteristics of pumpkin fruit.
研究领域:
植物学
·
食品科学
·
分析化学
✓ 已验证
产品关联
DOI
PubMed
高影响因子
LWT-FOOD SCIENCE AND TECHNOLOGY
·
2026
·
IF 7.1
Qun-Yan Xiong,Hui-Wen Gu,Qing-Qing Luo,Chao Wang,Yan-Qing Fu,Xiao-Li Yin
A machine learning-driven strategy for optimal feature selection is proposed. • W1W and W1C sensors serve as key variables for aroma type classification. • 131 volatile compounds were identified in Liupao tea. • 14 aroma type differentiating compounds were identified. • 4 key variables yielded 100% accurate aroma recognition models. This study aimed to differentiate between the stale-aroma and betelnut-aroma types of Liupao tea by analyzing their volatile profiles. An integrated methodology was employed, combining electronic nose (E-nose) technology with headspace gas chromatography-ion mobility spectrometry (HS-GC-IMS). The E-nose data effectively differentiated between the two aroma types in both dry leaves and tea infusions. HS-GC-IMS enabled identification of 131 volatile compounds. Key discriminative features were selected through machine learning by integrating the variable importance in projection (VIP) from orthogonal partial least squares discriminant analysis (OPLS-DA) with Shapley additive explanations (SHAP) values from extreme gradient boosting (XGBoost). This process identified two critical E-nose sensor channels and two key volatile compounds. Utilizing these four optimal variables, three classification models were developed: principal component analysis-linear discriminant analysis (PCA-LDA), partial least squares discriminant analysis (PLS-DA), and support vector machine (SVM). All models demonstrated a perfect classification accuracy of 100% on both the training and internal test datasets. The PCA-LDA and PLS-DA models maintained 100% accuracy on the independent test set, whereas the SVM model achieved a classification accuracy of 80% on the same set. These results reveal the volatile characteristics of Liupao tea and offer valuable insights for quality assessment and targeted flavor modulation.
✓ 已验证
DOI
高影响因子
FOOD RESEARCH INTERNATIONAL
·
2026
·
IF 8.8
Qi Zhao,Yicheng Zhou,Cong Wang,Mingxuan Li,Shan Zhu,Chang Xue,De Ji,Yuwen Qin,Chengxi Jiang
Gastrodia elata Bl. ( GE ), a widely utilized edible and functional material, exhibits notable quality variations due to distinct growing conditions across different geographical origins, making reliable methods for origin authentication and quality evaluation essential. This study integrated machine vision, flash GC e-nose and HPLC to analyze GE samples from Anhui (AH), Shaanxi (SN), and Yunnan (YN) provinces in China. Results showed comparable color characteristics among geographical origins, but distinct textural differences: SN samples displayed more complex textures with higher surface roughness, YN samples showed richer textural details but less distinct edges and lower periodicity, while AH samples exhibited smoother surfaces. Eleven flavor compounds were identified as potential flavor markers. Although no significant differences were found in the total content of six marker components, three linear machine learning models (PLS, SVM, and LDA) established based on multidimensional features achieved 100% classification accuracy in training and prediction sets, providing a reliable approach for GE geographical origin traceability.
✓ 已验证
产品关联
DOI
高影响因子
Analytical Methods
·
2026
·
IF 2.6
Jun-Xia Li,Min Gong,Chao Ma,Yao-Yao Li,Ling-Xiao Liu,Bin Wang,Jin-Jing Ren,Sheng-Ming Lei,Yun-Guo Liu
Rosa rugosa Thunb. can be skillfully blended with high-quality, mellow Baijiu to produce Rose Lujiu (RL), which emits a delightful aroma. In this study, e-nose, e-tongue, GC-MS, and GC-IMS were employed to analyze the aromatic compounds. Relative odor activity value (ROAV) analysis and orthogonal partial least squares discriminant analysis (OPLS-DA) were employed to comprehensively assess volatile compounds and predict key aroma compounds. A total of 111 volatile compounds were identified using GC-MS analysis. Additionally, GC-IMS analysis identified a total of 70 aroma compounds, among which 13 key contributors to the aroma were determined based on odor activity thresholds. These results indicate that ethyl acetate, ethyl butanoate, propyl hexanoate, and 3-methyl-1-butanol emerged as pivotal odorants enabling discrimination among different RL samples. These findings would provide valuable data for the application of commercial samples in the flavor modulation of RL.
✓ 已验证
产品关联
DOI
PubMed
高影响因子
FOOD RESEARCH INTERNATIONAL
·
2019
·
IF 3.579
Mostafa Gouda,Meihu Ma,Long Sheng,Xiaole Xiang
This study was focused on the differences between three concentrations of four safe bio-active volatile terpenes and natural compounds (trans-cinnamaldehyde, thymol, menthol, and vanillin) on the egg yolk volatile components. In which, Headspace Solid phase Micro Extraction (HS-SPME) followed by GC/MS were used for volatiles analysis, and electronic nose (E-Nose) with 18 sensors was used for the aroma pattern. In the results, a total of 111 different volatile compounds were identified by using GC/MS. Some ketones, amines, nitro and organic acid compounds such as hexanone were significantly reduced after using these compounds. On the other hand, first principal components (PC1) of trans-cinnamaldehyde (0.1%) was decreased significantly ( p < .05) to −4.8 ± 0.8, compared to the control with 4.3 ± 2.1. And, it increased benzaldehyde and nonanal peaks area significantly to 46.56 ± 14.46 and 9.22 ± 4.81 MPA. In menthol (0.1%), the content of d-limonene was decreased to 4.65 ± 0.49 MPA, compared to the control with 14.16 ± 4.95 MPA. On the other hand, by E-Nose, vanillin (0.1%) PC1 was decreased significantly to −4.48 ± 0.43. Moreover, thymol (0.1%) decreased hydrocarbon sensation to 0.448 ± 0.005. And, vanillin (0.1%) decreased hydrogen sulfide and ketones sensation to 0.169 ± 0.01 and 0.344 ± 0.01, respectively. In conclusion, this study is providing novel information to understand the effects of bio-active molecules on the biological fatty media volatility. Also, it explains how terpenes can protect egg yolk media from the unacceptable odor formation which results from the degradation of its lipids and proteins.
研究领域:
食品科学
·
分析化学
✓ 已验证
产品关联
DOI
PubMed
高影响因子
International Journal of Gastronomy and Food Science
·
2025
·
IF 3.6
Huiying Wu,Wensheng Yao,Shuangyu Ma,Dengyong Liu,Mingcheng Zhang
In this study, the evolution of key flavor compounds in sonit sheep lamb during grilling process has been studied by the E-nose, E-tongue, HS-GC-IMS technique and physicochemical methods. The results showed that the moisture content in lamb decreased, protein content increased due to water loss, fat content reduced due to thermal degradation during the grilling process. The pH showed an initial rapid decrease with significant differences at various grilling times (P < 0.05). The E-nose and E-tongue were found effective in distinguishing lamb samples during grilling duration. A total of 83 volatiles were identified by HS-GC-IMS. Among them, 29 flavor compounds with VIP > 1 were identified as the key flavor substances using the OPLS-DA model. Ethyl acrylate was the main characteristic flavor compounds in raw Sonit Sheep lamb, while ethanol and 1-butanol were determined as the key flavor compounds at 3 and 6 min. Ethylpyrazine, (E)-2-octenal, and nonanal are key substances at 9 min grilling. When grilled for 12 min, the flavor compounds such as 2-methylpropanal, ethyl ethanoate, and trimethylpyrazine became the most important flavor compound in the grilled lamb.
研究领域:
食品科学
·
分析化学
✓ 已验证
产品关联
DOI
高影响因子
LWT-FOOD SCIENCE AND TECHNOLOGY
·
2020
·
IF 4.006
Shuang Gu,Wei Chen,Zhenhe Wang,Jun Wang,Yujia Huo
This study described the rapid detection of milled rice infected with Aspergillus spp. species based on headspace-gas chromatography ion-mobility spectrometry (HS-GC-IMS) and electronic nose (E-nose) combined with chemometrics, namely principal component analysis (PCA), k-nearest neighbor (kNN) and partial least squares regression (PLSR). 3D HS-GC-IMS imaging and their response differences enabled the discrimination among fungal species. kNN was used to differentiate rice samples with cdifferent levels of fungal infection and achieved correct classified rate of 94.44% and 91.67% by HS-GC-IMS and E-nose, respectively. PLSR method was used for quantitative regression of fungal colony counts in rice samples and good prediction performances were achieved by HS-GC-IMS (R p 2 = 0.909, RMSEP = 0.202) and E-nose (R p 2 = 0.864, RMSEP = 0.235). The results indicated that both HS-GC-IMS and E-nose approaches can potentially be implemented for the detection of fungal contamination levels in milled rice, and HS-GC-IMS fingerprinting coupled with chemometrics might be used as an alternative tool for a highly sensitive method. This research might provide scientific information on the rapid, non-destructive, and effective fungal detection system for rice grains .
研究领域:
食品科学
·
微生物学
✓ 已验证
DOI
高影响因子
危险品化学品经营许可证（带存储）