The livestock meat and its nutrition quality is considered to be an important factor in our daily eating habits giving particular emphasis to health issues. The quality and the nutrition value of a raw-beef-steak, is highly connected with the fat percentage of it. Consequently, the determination of the fat percentage of a raw-beef-steak is crucial for meat producers and consumers as well. In this work, we present a fat mass estimation approach based on a state-of-the-art deep learning pipeline by utilizing a single colored image presenting raw-beef-steak. In order to produce more accurate outcomes, our pipeline combines two U-Nets, one for the background removal and one for the fat extraction. By following popular computational approaches we estimate the fat amount based on the pixels presenting it. To enhance the outcomes of this work, we introduce a new data-set annotated based on the needs of the experiment. The main goal of this work is to provide accurate nutritional information to end-users through novel technologies by exploiting a single image through a mobile application.
CITATION STYLE
Symeonidis, G., Kiourt, C., Kazakis, N. A., Nerantzis, E., & Nestor, T. (2022). Fat calculation from raw-beef-steak images through machine learning approaches: An end-to-end pipeline. In ACM International Conference Proceeding Series (pp. 110–115). Association for Computing Machinery. https://doi.org/10.1145/3575879.3575975
Mendeley helps you to discover research relevant for your work.