Naslov
Application of computer vision and image analysis method in cheese-quality evaluation: a review

Autori
Lukinac, Jasmina ; Jukić, Marko ; Mastanjević, Kristina ; Lučan, Mirela

Izvornik
Ukrainian food journal (2304-974X) 7 (2018), 2; 192-214

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, pregledni rad, znanstveni

Ključne riječi
computer vision ; image analysis ; cheese quality ; visual parameters

Sažetak
Introduction. This review paper deals with literature analysis of modern computer vision and image analysis method in cheese-quality evaluation. Materials and methods. In this paper several cheese types made from cow’s milk were analyzed: soft, semi-hard, and hard cheese respectively. All images shown in this paper were scanned with a flatbed scanner, and images were processed using ImageJ software. Results and discussion. Because with the most part of the external quality attributes evaluation is time-consuming, due to visual inspection, computer vision provides a means to perform this task automatically. To evaluate various cheese external quality properties (color), and defects (mechanical openings – gas holes ; rind defect – formation of calcium lactate crystals and excessive rind halo ; amount and distribution of added ingredients ; meltability and oiling-off) computer vision system has been successfully applied. Image analysis was used for measurement of the normal amount of gas production, and abnormal shape or distribution of eyes throughout the structure of Emmental and Tilsit cheese. Image analysis was used to determine the presence of rind halo cheese defect and to measure the area occupied by calcium lactate crystals on surfaces of naturally smoked Cheddar cheese samples. To improve empirical methods (Arnott and the Schreiber test) and to offer a new approach for evaluation of meltability and oiling-off of Mozzarella cheese, computer vision and image analysis method was applied. In addition, digital image analysis is used for evaluation of the addition of some ingredients and evaluation of the amount and distribution of added ingredients into the semi-cooked cheese with added paprika and herbs. Conclusion. Computer vision and digital image analysis represents an efficient and non-invasive technique able to investigate the cheese optical properties and give information concerning their composition and structure.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo, Prehrambena tehnologija

POVEZANOST RADA