Inspection of meat for trichinellosis. Classification of Trichinella spiralis using a convolutional neural network

3rd International Conference on Chemo and BioInformatics, Kragujevac, September 25-26. 2025. (pp. 355-363) 

 

АУТОР(И) / AUTHOR(S): Sofija Stanojlović

 

Download Full Pdf   

DOI:  10.46793/ICCBIKG25.355S

САЖЕТАК / ABSTRACT:

Trichinella spiralis is a dangerous parasite that lives in domestic and wild animals. In this paper, a method for the classification of this parasite was examined. This parasite most often lives in pigs and it is possible for it to enter the body by eating meat that has not been previously examined by a veterinarian. The paper aims to examine whether it is possible for a convolutional neural network to correctly classify images of healthy meat under a trichinoscope from images of infected meat under a trichinoscope. For this paper, a model was created that correctly classified the image of meat in 70% of cases. K-fold cross-validation was used for the process of training and evaluation of model performance. K-fold cross-validation divides the dataset into K equal folds. Then K iterations are executed. In each iteration, one of the folds is used as a validation set, while the remaining K-1 folds are used for model training. This technique reduces the risk of overfitting and is suitable for small datasets, such as the one in this paper. During the evaluation of the model, following the evaluation metrics, the following average results were obtained: accuracy of 0.85, recall of 0.80, precision of 0.89, and the F1-score value of 0.84. The model achieved good accuracy, but in order for the model to replace the work of a veterinarian, a higher accuracy should be achieved, which could potentially be obtained by using a larger dataset with a potentially deeper neural network architecture.

КЉУЧНЕ РЕЧИ / KEYWORDS:

Trichinella spiralis, parasite classification, convolutional neural network, K-fold cross- validation, meat inspection

ПРОЈЕКАТ / ACKNOWLEDGEMENT:

ЛИТЕРАТУРА / REFERENCES: