REPOSITORY > RESULTS

Abstract

Food fraud has been a persistent issue for decades, transforming from basic adulteration into a sophisticated, multibillion-dollar problem driven by significant economic incentives. With the global economic impact on the food industry estimated at USD 30–65 billion annually and over 10% of the world’s food supply affected each year, the consequences are far-reaching. Beyond financial losses, food fraud undermines consumer trust, damages the reputation of legitimate producers, and imposes heavy regulatory burdens, positioning fraud prevention as one of the most urgent and complex challenges confronting modern global food systems.
This doctoral dissertation addresses the critical need for strengthening food authenticity and traceability by introducing a novel, integrated framework comprising three complementary systems: a Physical Traceability System that relies on advanced analytical techniques and database-driven authentication; a Rapid Traceability System, enabling fast and reliable spectroscopic screening for fraud detection; and a Digital Traceability System, which employs cutting-edge Internet of Things (IoT) technologies to provide real-time, continuous monitoring across the entire supply chain.
The framework adopts a multi-disciplinary approach comprising (1) The application of stable isotope ratios and multi-elemental analyses to distinguish the geographical origin and production methods (conventional vs organic) of spices, (2) the use of FTIR spectroscopy to support the detection of adulteration and authenticity, (3) expanding the IsoFoodTrack authentication database by incorporating new isotopic, elemental, and metadata entries, (4) employing advanced chemometric techniques such as OPLS-DA and DD-SIMCA and AI models to interpret complex datasets, and (5) implementing IoT-enabled digital traceability for seafood supply chains through a pilot project in Montenegro. This approach is applied to four high-value commodities: saffron, paprika, cinnamon, and seafood, which were selected due to their high susceptibility to fraud.
The work yielded several key findings. The IsoFoodTrack database demonstrated a scalable and innovative framework for integrating isotopic, elemental, and related metadata to verify authenticity and ensure traceability. This architecture enabled the continuous inclusion of new analytical metrics and food categories, which strengthened its effectiveness in fraud detection over time. Authentic saffron samples achieved 94% classification accuracy for geographical-origin authentication using isotopic and elemental analyses, with DD-SIMCA identifying potential mislabeling in up to 15% of commercial samples, particularly those labelled as Iranian. For paprika and cinnamon, accuracies of 90% and 89%, respectively, were achieved. Differentiation by agricultural production methods was achieved with ~95% accuracy for cinnamon, while for paprika, characteristic isotopic (δ15N, δ34S) and elemental (Zn) markers enabled reliable separation of production types. FTIR spectroscopy further distinguished Ceylon from cassia cinnamon and highlighted possible paprika adulteration. Applying artificial intelligence modeling, specifically the Light Gradient Boosting Machine (LightGBM), to 115 saffron samples achieved an overall classification accuracy of 93% using combined isotopic and elemental data. SHapley Additive exPlanations (SHAP) analysis identified the most influential elemental discriminators Ni, Co, K, Mn, Sr, Rb, Ga, Al, Cd, and Pb, enabling a targeted analytical approach that could reduce analysis costs by approximately 79% while maintaining classification accuracies above 90%. The IoT-enabled system was designed to facilitate real-time monitoring of seafood supply chains by integrating environmental sensors (water temperature, dissolved oxygen, pH, and salinity) with GPS and onboard sensors for transport conditions, all managed via Cloud-based data. QR codes will enable interactions with users and stakeholders outside the fish farm, providing product/package information and collecting their feedback.
The integrated traceability framework presented in this thesis provides a highly effective approach for authenticating and tracing high-value commodities such as saffron, paprika, cinnamon, and seafood. Through its advanced capabilities in detecting geographical origins, production methods, and adulteration, it proves a powerful tool in combating food fraud. Future work should broaden geographical and commodity scope via international collaborations and open-data platforms, incorporate explainable AI (XAI) for transparent decisions, and combine compound-specific isotope analysis (CSIA) with metabolomics to further improve analytical precision. Policymakers should promote AI, machine learning, and blockchain integration into existing traceability systems to boost transparency and fraud prevention.