Evaluation of an artificial olfactory system for grain quality discrimination

This study was conducted to evaluate the performance of a commercial electronic nose (Cyranose 320™) for sensing indicator compounds (ethanol and acetic acid) associated with spoiled beef. The present study reported the sensitivity of an array of 32 sensors to ethanol and acetic acid. Different vapor concentrations of ethanol (37, 100, 250 and 500 ppm) and acetic acid (38, 75, 100, 200 and 300 ppm) were tested to evaluate the performance of the commercial system. An in-house designed universal gas sensing and characterization system was coupled with the electronic nose system to generate the desired gas concentration. The raw smell print patterns were obtained and analyzed for individual and multiple detectors. Tukey’s multiple comparison technique was performed to analyze the response of individual detectors. Area above and below the baseline were selected as two features for pair wise comparison of the detectors. Different sensors showed different responses between various concentrations of gases. Analysis of multiple detectors was performed using linear and quadratic discriminant analysis (LDA and QDA) along with bootstrap. LDA along with bootstrap provided the highest total classification accuracies of 94.34 % between 100 and 200 ppm of acetic acid. QDA provided higher total classification accuracy of 89.69 % at lower concentration level of 38 and 75 ppm for acetic acid and 84.78 % between 37 and 100 ppm of ethanol. Hence, QDA was a better model of choice at lower concentrations. This study proved non-selective nature of sensors and showed that simultaneous use of multiple sensor information provided better classification accuracy for discriminating various gas concentrations as compared to using individual sensor output.

The aim of this study was to develop a portable electronic nose system for freshness measurement of stored pork. An electronic nose system was constructed using seven different MOS sensor array. To determine the quality change of pork with storage time, the samples were divided into ten groups in terms of storage time with an increment of 2 day up to 19 storage days. GC-MS, total bacteria's count (TBC), thiobarbituric acid reactive substance (TBARS), and pH analyses as well as the analysis of the electronic nose system measurement were performed to monitor the freshness change of the samples. To investigate the performance of the electronic nose system for detecting the change of freshness of pork, the acquired signal values of the system were compared with those of GC-MS, TBC, TBARS, and pH analysis values. According to principal component analysis (PCA) and linear discriminant analysis (LDA) with the signals of the electronic nose system for the pork samples, the sample groups were clearly separated into two groups of 1-9 days and 11-19 days, and four groups of 1-3 days, 5-9 days, 11 days, and 13-19 days respectively. The results show that the electronic nose system has potential for evaluating freshness of pork.

The aim of the present study was to develop an electronic nose for the quality control of red meat. Electronic nose and bacteriological measurements are performed to analyse samples of beef and sheep meat stored at 4°C for up to 15 days. Principal component analysis (PCA) and support vector machine (SVM) based classification techniques are used to investigate the performance of the electronic nose system in the spoilage classification of red meats. The bacteriological method was selected as the reference method to consistently train the electronic nose system. The SVM models built classified meat samples based on the total microbial population into "unspoiled" (microbial counts < 6 log10 cfu/g) and "spoiled" (microbial counts ≥ 6 log10 cfu/g). The preliminary results obtained by the bacteria total viable counts (TVC) show that the shelf-life of beef and sheep meats stored at 4 °C are 7 and 5 days, respectively. The electronic nose system coupled to SVM could discriminate between unspoiled/ spoiled beef or sheep meats with a success rate of 98.81 or 96.43 %, respectively. To investigate whether the results of the electronic nose correlated well with the results of the bacteriological analysis, partial least squares (PLS) calibration models were built and validated. Good correlation coefficients between the electronic nose signals and bacteriological data were obtained, a clear indication that the electronic nose system can become a simple and rapid technique for the quality control of red meats.

Neural-network-integrated electronic nose system for identification of spoiled beef

We evaluated the directional asymmetry between right and left wings and quantified the intraspecific and interspecific variation of the wing shape of 601 specimens of the genus Diachlorus to determine to what extent the geometrical variation discriminates six species distributed in six protected areas of Colombia. Geometric analyses were performed, integrating Procrustes methods, principal component analyses, cluster analyses, linear and quadratic discriminant analyses, and evaluations of shape changes. In Diachlorus, left and right wings did not present significant asymmetry but a geometrical analysis was allowed for species identification and, in some cases, the origin of the specimens using the variation of wing shape; the best-assigned species was Diachlorus leticia Wilkerson & Fairchild, while the worst was Diachlorus jobbinsi Fairchild, which also had the highest intraspecific variation, while Diachlorus fuscistigma Lutz had the lowest variation. Diachlorus fuscistigma and Diachlorus leucotibialis Wilkerson & Fairchild were the most similar species, while D. leucotibialis and Diachlorus nuneztovari Fairchild & Ortiz were the most disimilar. The specimens with the most different wing shape belonged to Chocó (especially those of D. jobbinsi), the geographically farthest area from the others in the study; however, no correlation was observed between geometric and geographical distances. Linear discriminants were better than nonlinear (quadratic) discriminant analyses in predicting species membership, but the opposite was true for predicting area membership. Based on our data, we hypothesized that other species of Diachlorus could also be discriminated using geometric morphometry of the wing shape.

Aminoleban® EN is a bitter total enteral nutrient product with an unpleasant odour. The purpose of this study was to evaluate the odour of Aminoleban® EN, taste-masked with various flavoured powders, by human testing and by using an electronic nose system (Alpha M.O.S.). Initially, the odour of Aminoleban® EN alone at various concentrations was evaluated. Secondly, the odour of Aminoleban® EN with the addition of various flavoured powders was evaluated; the evaluation included data on odour intensity, preference levels and evoked taste image. With the human nose, a concentration of 0.43 g/100 ml seemed to be the taste threshold for Aminoleban® EN. In the electronic nose system, principal component analysis revealed that concentrations over 0.39 g/100 ml Aminoleban® EN could be distinguished from water. It was clear from both human testing and using the electronic nose system that the unpleasant odour of Aminoleban® EN could be masked by pineapple- and plum-flavoured powders, both of which contain not only sweet but also sour components. The electronic nose system may be useful for detecting odours at or below the level detectable by the human olfactory system and for classifying the quality of different odours, thus reducing the amount of human testing required.

The performance of seven classification methods, binary logistic (BLR), probit (PR) and cumulative probit (CPR) regression, linear (LDA) and quadratic (QDA) discriminant analysis, artificial neural networks (ANN), and naïve Bayes classification (NBC), is examined in skeletal sex estimation. These methods were tested using cranial and pelvic sexually dimorphic traits recorded on a modern documented collection, the Athens Collection. For their implementation, an R package has been written to perform cross-validated (CV) sex classification and give the discriminant function of each of the methods studied. A simple algorithm that combines two discriminant functions is also proposed. It was found that the differences in the classification performance between BLR, PR, CPR, LDA, QDA, ANN, and NBC are overall small. However, LDA is simpler and more flexible than CPR, QDA, and ANN and has a small but clear advantage over BLR, NBC, and PR. Consequently, LDA may be preferred in skeletal sex estimation. Finally, it is striking that the combination of pelvic and cranial traits via their discriminant functions, determined either by BLR or LDA, removes practically any population-specificity and yields much better predictions than the individual functions; in fact, the prediction accuracy increases above 97%.

PSCFormer: A lightweight hybrid network for gas identification in electronic nose system

Crop storage should be carried out under hygienic conditions to ensure safe products, but sometimes grain dust which has settled from previous storage may be left over and incorporated to the following stored grains. This paper describes the results obtained using a lab model developed in order to assess the impact of grain dust incorporation for its direct contribution as a contaminant but also as an inoculum in stored wheat. Settled grain dust (4 samples) released from Belgian grain storages were collected and analysed by HPLC for ergosterol, ochratoxin A (OTA) and citrinin (CIT) content. For OTA and for ergosterol, there was a high degree of variability in concentrations found in the dust samples (from 17.3–318 ng g−1 and from 39–823 µg g−1, respectively) whilst for CIT, the range was less significant (from 137–344 ng g−1). Incorporation of grain dust into wheat storage contributed to an increase in the concentrations of mycotoxins in the stored grain. Dust acts as a contaminant and as an inoculum. According to these two ways, patterns of mycotoxin generation vary with the nature of the mycotoxin, the mycotoxigenic potential of dust and the water activity of the wheat. OTA and CIT showed a very versatile image when considering the amounts of toxins produced under the selected experimental conditions. The development of a robust tool to forecast the mycotoxigenicity of dust was based on the determination of ergosterol content as a general marker of fungal biomass. Present results suggest that this predictive tool would only be valid for predicting the contamination level of CIT and OTA at reasonable moisture content (14–20%). The potential risk of having highly contaminated batches from stock to stock may thus occur and this paper discusses possible pathways leading to OTA and CIT contamination either under wet or dry storage conditions. We therefore, recommend taking precautionary measures not only by controlling and maintaining moisture at a reasonable level during storage of the raw materials but also by paying more attention to the cleaning of the stores before loading in the new harvests.

Detection of terbufos in cases of intoxication by means of entomotoxicological analysis using ATR-FTIR spectroscopy combined with chemometrics

Electronic nose (EN) system based on an array of gas sensors and its recognition techniques have been widely investigated in recent decades. However, the integration of a portable EN system is not mature yet. We present an integrated EN system based on an array of gas sensor fabricated in our in-house foundry. A data acquisition technique based on a resistance-to-time conversion is proposed enabling to remove the need for an analog-to-digital converter enabling to reduce the power consumption and area. This ADC-free system is implemented in a multi-chip platform and enabling gas sensing and recognition functionalities. This system is compact, low cost and low power which is suitable for portable applications. Both sensing and recognition operations have been successfully demonstrated in our demo system.

Ultraviolet (UV) hyperspectral imaging shows significant promise for the classification and quality assessment of raw cotton, a key material in the textile industry. This study evaluates the efficacy of UV hyperspectral imaging (225–408 nm) using two different light sources: xenon arc (XBO) and deuterium lamps, in comparison to NIR hyperspectral imaging. The aim is to determine which light source provides better differentiation between cotton types in UV hyperspectral imaging, as each interacts differently with the materials, potentially affecting imaging quality and classification accuracy. Principal component analysis (PCA) and Quadratic Discriminant Analysis (QDA) were employed to differentiate between various cotton types and hemp plant. PCA for the XBO illumination revealed that the first three principal components (PCs) accounted for 94.8% of the total variance: PC1 (78.4%) and PC2 (11.6%) clustered the samples into four main groups—hemp (HP), recycled cotton (RcC), and organic cotton (OC) from the other cotton samples—while PC3 (6%) further separated RcC. When using the deuterium light source, the first three PCs explained 89.4% of the variance, effectively distinguishing sample types such as HP, RcC, and OC from the remaining samples, with PC3 clearly separating RcC. When combining the PCA scores with QDA, the classification accuracy reached 76.1% for the XBO light source and 85.1% for the deuterium light source. Furthermore, a deep learning technique called a fully connected neural network for classification was applied. The classification accuracy for the XBO and deuterium light sources reached 83.6% and 90.1%, respectively. The results highlight the ability of this method to differentiate conventional and organic cotton, as well as hemp, and to identify distinct types of recycled cotton, suggesting varying recycling processes and possible common origins with raw cotton. These findings underscore the potential of UV hyperspectral imaging, coupled with chemometric models, as a powerful tool for enhancing cotton classification accuracy in the textile industry.

Objective: Delirium is an important syndrome found in patients in the intensive care unit (ICU), however, it is usually under-recognized during treatment. This study was performed to investigate whether delirious patients can be successfully distinguished from non-delirious patients by using heart rate variability (HRV) and machine learning. Approach: Electrocardiography data of 140 patients was acquired during daily ICU care, and HRV data were analyzed. Delirium, including its type, severity, and etiologies, was evaluated daily by trained psychiatrists. HRV data and various machine learning algorithms including linear support vector machine (SVM), SVM with radial basis function (RBF) kernels, linear extreme learning machine (ELM), ELM with RBF kernels, linear discriminant analysis, and quadratic discriminant analysis were utilized to distinguish delirium patients from non-delirium patients. Main results: HRV data of 4797 ECGs were included, and 39 patients had delirium at least once during their ICU stay. The maximum classification accuracy was acquired using SVM with RBF kernels. Our prediction method based on HRV with machine learning was comparable to previous delirium prediction models using massive amounts of clinical information. Significance: Our results show that autonomic alterations could be a significant feature of patients with delirium in the ICU, suggesting the potential for the automatic prediction and early detection of delirium based on HRV with machine learning.

An electronic nose system, which imitates the function of the human olfactory system, has been developed and applied to evaluate the quality of cigarettes. Two major parts of the system, that is, the sensor array and the neural network, are presented. The system has been used to evaluate and classify five different cigarette smokes. The feasibility of evaluating the quality of cigarettes with the electronic nose system has proved sound by the preliminary experiment and the resulting data.

We investigated three classifiers for the task of distinguishing between benign and malignant breast lesions. Classification performance was measured in terms of area under the ROC curve (AUC value). We compared linear discriminant analysis (LDA), quadratic discriminant analysis (QDA) and a Bayesian neural net (BNN) with 5 hidden units. For each lesion, 46 image features were extracted and principal component analysis (PCA) of these features was used as classifier input. For each classifier, the optimal number of principal components was determined by performing PCA within each step of a leave-one-case-out protocol for the training dataset (1125 lesions, 14% cancer prevalence) and determining which number of components maximized the AUC value. Subsequently, each classifier was trained on the training dataset and applied ‘cold turkey’ to an independent test set from a different population (341 lesions, 30% cancer prevalence). The optimal number of principal components for LDA was 24, accounting for 97% of the variance in the image features. For QDA and BNN, these numbers were 5 (70%) and 15 (93%), respectively. The LDA, QDA and BNN obtained AUC values of 0.88, 0.85, and 0.91, respectively, in the leave-one-case-out analysis. In the independent test – with AUCs of 0.88, 0.76, and 0.82 – only LDA achieved performance identical to that for the training set (lower bound of 95% non-inferiority interval -.0067), while the others performed significantly worse (p-values << 0.05). While the more complex BNN classifier outperformed the others in leave-one-case-out of a large dataset, LDA was the robust best-performer in an independent test.