Instrumental Data Fusion for Food Analysis Application

Nachev, Veselin; Titova, Tanya; Kosturkov, Rosen

doi:10.1016/j.ifacol.2019.12.446

Cited by 3 publications

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Discriminant analysis. The discriminant analysis is a data classification method that uses a grouping variable, as explained by Kirilova et al [17] and Nachev et al [18]. The procedure can be implemented through the use of linear or non-linear partition functions.…”

Section: Classification Methodsmentioning

confidence: 99%

Integration of Sensor Fusion to Enhance Quality Assessment of White Brine Cheeses

Zlatev,

Spahiu,

Taneva

et al. 2024

Dairy

View full text Add to dashboard Cite

The article examines the main characteristics of white brine cheeses from different manufacturers and changes in their quality indicators. These characteristics include the active acidity, electrical conductivity, total dissolved solids, oxidation–reduction potential, and organoleptic assessment. In this context, the connection to biomimetics lies in the approach of integrating multiple sensory modalities, similar to how biological systems often use multiple senses to perceive and understand their environment. For this purpose, spectral, ultrasonic, and gas characteristics were used, from which informative indices were extracted, united at a later stage in a vector of features. Based on the classification, it was found that the optical characteristics of cheeses from different manufacturers overlap, thus making it possible to predict the main indicators for each type of cheese. The results show that the use of a multimodal approach combining features from different sensors contributes to a better understanding of the variations in cheese properties, while improving the predictive abilities of the created models. The obtained results give a clear idea of the quality of the cheese, thus enabling adequate decisions to be made during the production process.

show abstract

Section: Classification Methodsmentioning

confidence: 99%