Nondestructive assessment of woody breast myopathy in chicken fillets using optical coherence tomography imaging with machine learning: a feasibility study

Authors: EKRAMIRAD, NADER - Oak Ridge Institute For Science And Education (ORISE); Yoon, Seung-Chul; Bowker, Brian; Zhuang, Hong

Submitted to: Food and Bioprocess Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/10/2024
Publication Date: 3/19/2024
Citation: Ekramirad, N., Yoon, S.C., Bowker, B.C., Zhuang, H. 2024. Nondestructive assessment of woody breast myopathy in chicken fillets using optical coherence tomography imaging with machine learning: a feasibility study. Food and Bioprocess Technology. https://doi.org/10.1007/s11947-024-03369-1.
DOI: https://doi.org/10.1007/s11947-024-03369-1

Interpretive Summary: Woody breast (WB) myopathy is a significant muscular abnormality that leads to the development of excessively hardened chicken fillets. The severe condition of woody breast can have detrimental effects on the poultry industry, including reduced meat quality, increased waste, degradation of nutritional content, and diminished customer satisfaction, resulting in substantial economic losses. The current gold standard for characterizing the physical properties of muscle tissue affected by woody breast is a histological technique that employs a light microscope. However, this approach is destructive, expensive, time-consuming, and limited to analyzing small sample areas. Consequently, the objective and rapid assessment of the extent of WB myopathy in individual fillets poses considerable challenges. In this study, optical coherence tomography (OCT) combined with image processing and machine learning methods was employed for the detection and classification of WB chicken fillets. The study provided the potential of OCT as a sub-surface microstructure imaging method for assessing chicken fillets for the severity of WB condition. Since the distribution of WB areas along a chicken fillet is heterogenous, in this study, the OCT was adapted for large-scale scanning along a whole fillet using a semi-automated imaging system. The acquired images for each fillet were stitched together to be analyzed using an automated image processing algorithm to extract relevant features to be used to train machine learning models to classify or predict WB condition of the fillet. The results showed an excellent performance in the classification of normal and severe WB samples. Additionally, the classification of normal from moderate and severe WB samples combined achieved a test accuracy of 93.3%. However, the accuracy for the classification of normal, moderate, and severe WB samples were found to be lower with a maximum test accuracy of 80%. Overall, the results of this study suggest that the OCT imaging coupled with image processing and machine learning techniques can enable the development of an automated high-speed scanning system as an effective non-invasive method for evaluating WB in chicken meat, and potentially for other agricultural and horticultural products.

Technical Abstract: Woody breast (WB) myopathy is a major muscle abnormality, causing excessive hardness and chewiness in chicken fillets. The woody breast condition can potentially cause big economical losses in the poultry industry by decreasing meat quality, increasing waste, degrading nutritional content, and reducing customer satisfaction. A histological technique using a light microscope has been the gold standard to characterize the sub-surface properties of the muscle with the woody breast condition, which is destructive, costly, time-consuming, and limited to analyzing only small sample areas. It is currently very challenging to assess the degree of WB myopathy objectively and rapidly in individual fillets. There is a need to develop an effective sensing technology for rapidly characterizing the woody breast condition by measuring the sub-surface cross-sections of the entire fillet at a high resolution. In this study, optical coherence tomography (OCT) was adapted to image the sub-surface microstructure of chicken muscle tissue along an entire fillet at a micrometer resolution. The extracted features from OCT images were analyzed to be used as input to machine learning models to classify chicken fillets based on the WB severity. The results demonstrated excellent classification performance in the two-class task of normal vs. severe WB samples. Additionally, the task of distinguishing normal from moderate and severe WB samples combined achieved a test accuracy of 93.3%. The accuracies for the three-class classification of normal, moderate, and severe WB samples were lower compared to the rates achieved in the two-class scenarios. Overall, the large-scale OCT imaging, showed to be an effective non-invasive method for evaluating WB in chicken meat, and potentially for other agricultural products.