Метод пошуку і аналізу Е-домішок та інших складників у продуктах харчування населення

Abstract

The research is dedicated to the development of a formal method and a corresponding intelligent IT-system that allows consumers to automatically determine the content of food additives (E-additives) and provide an assessment of potential health risks based on EFSA and WHO data by photographing a product label. To implement the proposed approach, a combination of Natural Language Processing (NLP) methods for label text analysis, Computer Vision (CV) for ingredient recognition, and Machine Learning (ML) for classifying their hazard based on EFSA and WHO data was used. The experimental results showed that the system achieved an accuracy of 94% in recognizing E-additives on the test dataset (10,000 images). It was found that 23% of the analyzed products contain additives with potential allergenicity (for example, E320, E621). Furthermore, highly processed products contain a relatively larger number of additives, which is fully consistent with the results of previous studies in the field of food toxicology. The proposed method and the technological means for its implementation are promising for mass monitoring of food quality and consumer informing.