News

This study (open access) builds on a previously-published proof of concept.  The authors are working towards producing a hand-held multi-mode scanner (combining fluorescence, visible, NIR, and short-wave IR spectroscopy) to support species verification of white fish fillets in business-to-business supply (currently reliant, largely, on visual recognition by experienced traders).

The explain that one of the key challenges in using machine learning for fish species identification is managing the large number of classes, as the variety of fish species is extensive. In their previous research, they introduced a novel multi-mode, highly multi-class machine learning framework based on a hierarchy of dispute models. This approach involved training a global model, and then recognizing groups of classes that have feature subspaces too similar for effective single-stage classification. By partitioning the overall space into smaller, distinct subspaces, they trained specialized models that are more tailored to these specific subsets of the dataset. In practice, the global model initially classified a sample to determine the appropriate subspace, while the dispute model then identified the precise species within that subspace.

The objective of this latest study was to apply this approach data acquired with the multi-mode handheld spectroscopy device. Tissue spectra were acquired at 25 positions on 68 fillets from 11 species, in both frozen and thawed states.

They report that feature-level fusion across the four spectroscopy modes enabled higher classification accuracy than any single mode alone. A global machine-learning model classified all species with 85 ± 2.8 %, while specialized dispute models for commonly misclassified species improved performance to 90 % ± 6.1 %. Individual models for thawed and frozen fillets achieved 90 ± 6.0 % and 90 ± 5.4 %, respectively, with dispute models in the thawed dataset increasing accuracy to 93 ± 4.3 %.

They conclude that their results demonstrate that portable multi-mode spectroscopy, combined with machine learning, can provide a fast, non-destructive and reliable tool for on-site fish species identification.

This paper (open access) reports the development of a classification model for the geographic origin on black tea based upon measuring a panel of 15 trace elements by X-ray fluorescence (XRF).  XRF is a non-destructive technique.  The only sample preparation required is grinding the tea leaves into a fine powder.

The model could discriminate between 10 major tea-producing regions.  It was built using reference samples obtained, via tea industry contacts, directly from plantations or primary processing facilities.   791 black tea samples were collected in total: Assam (272 samples), Burundi (40 samples), Darjeeling (145 samples), Ethiopia (40 samples), Keemun (115 samples), Kenya region 1 (41 samples), Kenya region 2 (40 samples), Malawi (40 samples), Rwanda (10 samples), and Sri Lanka (48 samples).

Two unsupervised analysis techniques were used to visualize high-dimensional data, and six supervised models were employed to discriminate the ten GI regions.

The authors report that machine learning models, including random forest, support vector machine, k-nearest neighbours, linear discriminate analysis, and the deep learning multilayer perceptron (MLP) model, demonstrated superior predictive capabilities compared to the traditional partial least squares discriminant analysis model. The MLP model achieved the highest performance, with a 97.7 % overall F1 score in predicting the geographical origins of 532 authentic samples across ten GI regions.

The authors also Identified Rb, Sr, Mn, Si, and Cl as geographical markers for African region discrimination.

The conclude that their work could form the basis and foundation for an international database of tea Geographic Origin, enabling cheap and quick authenticity verification testing.

Photo by Oleg Guijinsky on Unsplash

Classification models for food authenticity tests can - in principle -  be based on any analytical technique that collects multi-variate data.  In the case of spectrometric data (such as NIR or multi-spectral imaging) the equipment can be relatively cheap.  For collecting chemical data, researchers often use high-end equipment such as advanced LC-MS or GC-MS

This proof-of-concept study (purchase required) is a rare example of building a classification model using a cheaper test (HPLC with fluorescence detection) to measure a chemical parameter.  The authors prepared cold-pressed walnut and pumpkin seed oils adulterated with 0 – 50% of sunflower oil.  They developed a classification model based on the concentrations of the four tocopherols (α-, β-, γ-, and δ-).  They report that the model was capable of discriminating sunflower oil adulteration down to 2-3%.

This paper (open access) provides a comprehensive overview of emerging non-invasive techniques—such as fluorescence, near-infrared, mid-infrared, and Raman spectroscopy—for assessing meat quality and detecting adulteration.

The key novelty of this review is its integration of bibliometric analysis with a critical evaluation of advanced technologies aligned with the UN Sustainable Development Goals. Within the tabulated lists of published papers, the authors add their own 1-line opinion on the robustness of the underpinning database or chemometrics, and how near the work is to practical application.

The review highlights the potential of hybrid systems that integrate spectroscopy with chemometrics and machine learning to provide accurate, real-time, and sustainable meat authentication solutions. It also highlights research gaps such as the need for multi-adulterant detection models, standardized validation protocols, and open-access spectral databases.

The authors aim to align their commentary on innovation with regulatory and sustainability frameworks, including the UN Sustainable Development Goals.

Photo by Victoria Shes on Unsplash

Fluorescence spectroscopy utilizing benchtop and portable spectrometers with light-emitting diodes (LEDs) as a fixed excitation source has been used as a method for detecting food adulteration in various products, including honey, extra virgin olive oil, tea, and coffee  It is cost-effective, rapid, and sensitive, allowing for intact measurement. LED-based fluorescence spectroscopy is fast, accurate, and cheaper than using a laser.. Recent advancements in semiconductor technology have enabled the delivery of LEDs with commercially available wavelengths ranging from 370 to 470 nm, exhibiting significant light intensity.

In this paper (purchase required), the authors used the technique to develop a classification model to detect ground soy in ground-roasted Arabica coffee, and to differentiate Robusta and Liberica varieties.  The abstract gives no details of the reference samples used to construct or validate the model but it was limited to 2024 season samples harvested in Indonesia.

Photo by Nathan Dumlao on Unsplash

Measuring the profile of trace metals in food is one approach to discriminating its geographic origin.  Analytical tools for trace-metal measurement tend to be laboratory-based with expensive capital equipment.

There has been recent research to develop point-of-use sensors for metal ions using specific chemical binders (often based on the chemistry of human olefaction) with fluorescent markers.  This paper (purchase requires) takes the work a significant step forward.  Their sensor is based upon the principle that different metal ions induce different degrees of aggregation in perylene diimide derivative based supramolecular nanoaggregates.  This enabled the construction of a multi-analyte sensor which they report as having ease of preparation, rapid response, and high sensitivity originating from large specific surface areas.  The authors report that they used the sensor to build a successful classification model of geographic origin for both drinking water and apples.