Determination of Biodegradability of Plant and Animal Based Edible Film Packaging: Artificial Intelligence Application
DOI:
https://doi.org/10.24925/turjaf.v14i4.1068-1077.8523Keywords:
Artificial Intelligence , Deep Learning , Renewable Film , BiodegradabilityAbstract
This study aims to by classifying the direct biodegradability status (day 0 and day 30) of edible films (plant-based Pectin, and animal-based Whey Protein Isolate-WPI) through visual data. For this purpose, a unique dataset consisting of 399 images was created using a controlled soil-burial experiment with four classes (Pectin 0, Pectin 30, WPI 0, WPI 30). Using this dataset, the performance of four different Convolutional Neural Network (CNN) architectures (ResNet-50, DenseNet-121, Xception, and VGG-16) was comparatively analysed using transfer learning and data augmentation strategies. Experimental results have demonstrated that the DenseNet-121 architecture provides a clear advantage over all other models. DenseNet-121 achieved 98,36% accuracy, 98,44% precision, and 98,33% F1-score, misclassifying only one of the 59 examples in the test set. Notably, this model, which demonstrated the highest performance, also proved to be the most efficient architecture with 6,96 million parameters; in contrast, ResNet-50 exhibited the lowest performance (93,44% accuracy). These findings demonstrate that DenseNet’s dense feature reuse philosophy is highly effective in capturing subtle morphological differences between degraded and undegraded films. These results provide a concrete foundation for developing a rapid, non-invasive, and low-cost automated system for the biodegradability certification of renewable films. These findings demonstrate that DenseNet’s dense feature reuse philosophy is highly effective in capturing subtle morphological differences between degraded and undegraded films. Additionally, the Grad-CAM (Gradient-weighted Class Activation Mapping) analysis, performed visually, demonstrated that the model focused directly on morphological deformations and structural distortions on the film surface rather than on background noise when making decisions. This study provides a concrete foundation for developing a rapid, non-invasive, and low-cost automated system for the biodegradability certification of renewable films.
References
Abookleesh, F., Zubair, M., & Ullah, A. (2025). Leveraging machine learning for the optimization of reinforced rapeseed protein-gelatin edible coatings for enhanced food preservation. Chemical Engineering Journal, 512, 162604. https://doi.org/10.1016/j.cej.2025.162604
Chaichi, M., Hashemi, M., Badii, F., & Mohammadi, A. (2017). Preparation and characterization of a novel bionanocomposite edible film based on pectin and crystalline nanocellulose. Carbohydrate Polymers, 157, 167–175. https://doi.org/10.1016/j.carbpol.2016.09.062
Chollet, F. (2017). Xception: Deep Learning with Depthwise Separable Convolutions.
e Silva, N. da S., de Souza Farias, F., dos Santos Freitas, M. M., Hernández, E. J. G. P., Dantas, V. V., Oliveira, M. E. C., Joele, M. R. S. P., & Lourenço, L. de F. H. (2021). Artificial intelligence application for classification and selection of fish gelatin packaging film produced with incorporation of palm oil and plant essential oils. Food Packaging and Shelf Life, 27, 100611. https://doi.org/10.3390/ijms26146975
Fredi, G., & Pegoretti, A. (2024). Yeşil Bir Gelecek için Biyobozunur Plastikler (Y. Öztekin & Ü. Sayın, Eds.).
Işık, H., Dağhan, Ş., & Gökmen, S. (2013). Gıda endüstrisinde kullanılan yenilebilir kaplamalar üzerine bir araştırma. Gıda Teknolojileri Elektronik Dergi, 8(1), 26-35.
Gong, W., Yao, H. Bin, Chen, T., Xu, Y., Fang, Y., Zhang, H. Y., Li, B. W., & Hu, J. N. (2023). Smartphone platform based on gelatin methacryloyl(GelMA)combined with deep learning models for real-time monitoring of food freshness. Talanta, 253. https://doi.org/10.1016/j.talanta.2022.124057
Gupta, V., Biswas, D., & Roy, S. (2022). A comprehensive review of biodegradable polymer-based films and coatings and their food packaging applications. Materials, 15(17), 5899. https://doi.org/10.3390/ma15175899
He, K., Zhang, X., Ren, S., & Sun, J. (n.d.). Deep Residual Learning for Image Recognition. Retrieved September 25, 2025, from http://image-net.org/challenges/LSVRC/2015/
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (n.d.). Densely Connected Convolutional Networks. Retrieved September 25, 2025, from https://github.com/liuzhuang13/DenseNet.
Jaramillo, C. M., Gutiérrez, T. J., Goyanes, S., Bernal, C., & Famá, L. (2016). Biodegradability and plasticizing effect of yerba mate extract on cassava starch edible films. Carbohydrate Polymers, 151, 150–159.
Javanmardi, S., & Ashtiani, S. H. M. (2025). AI-driven deep learning framework for shelf life prediction of edible mushrooms. Postharvest Biology and Technology, 222. https://doi.org/10.1016/j.postharvbio.2025.113396
Kavrut, E. (2024). Production of pH-Sensitive natural decomposition indicator with whey protein isolate-based edible film to monitor fish freshness. Latin Am Appl Res, 54(2), 247–252. DOI: 10.52292/j.laar.2024.3098
Kavrut, E., & Sezer, Ç. (2024). The antimicrobial effect of edible film packaging in instant meatballs on Escherichia coli O157: H7. Journal of the Hellenic Veterinary Medical Society, 75(1), 7059–7072. doi: 10.12681/jhvms.34084
Kavrut, E., Sezer, Ç., & Alwazeer, D. (2024). A bibliometric analysis: what do we know about edible coatings? Journal of Food Science and Technology, 61(11), 2057–2070.
Luo, Q., Rong, X., Xiao, Z., Duan, X., Zhou, Y., Zhang, J., Wang, X., Peng, Z., Dai, J., Liu, Y., & Fang, Z. (2025). Effect of chitosan films containing clove essential oil-loaded microemulsions combined with deep learning on pork preservation and freshness monitoring. Food Control, 168. https://doi.org/10.1016/j.foodcont.2024.110914
Niam, S., Ahmad, I., Rayhan, M. A., Mahmood, S., Jon, P. H., & Ahmed, M. M. (2025). Machine learning-based optimization of alginate, guar gum, and pectin-based edible coatings for extended strawberry shelf life. LWT, 233. https://doi.org/10.1016/j.lwt.2025.118548
Örücü, S., & Gökmen, S. (2023). Makine Öğrenme metotlarının mantı kalitesinin belirlenmesinde kullanılabilirliği. Uluslararası İleri Doğa Bilimleri ve Mühendislik Araştırmaları Dergisi 7(10), 175-181
Riaz, A., Lagnika, C., Luo, H., Dai, Z., Nie, M., Hashim, M. M., ... & Li, D. (2020). Chitosan-based biodegradable active food packaging film containing Chinese chive (Allium tuberosum) root extract for food application. International Journal of Biological Macromolecules, 150, 595-604.
Silue, Y., & Fawole, O. A. (2024). Global Research Network Analysis of Edible Coatings and Films for Preserving Perishable Fruit Crops: Current Status and Future Directions. In Foods (Vol. 13, Issue 15). Multidisciplinary Digital Publishing Institute (MDPI). https://doi.org/10.3390/foods13152321
Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition. http://www.robots.ox.ac.uk/
Wang, J., Xia, L., Liu, H., Zhao, C., Ming, S., & Wu, J. (2024). Colorimetric microneedle sensor using deep learning algorithm for meat freshness monitoring. Chemical Engineering Journal, 481. https://doi.org/10.1016/j.cej.2023.148474
Waqas, M., Chen, Z., Abbas, Y., Farooq, A., Han, X., Zhong, H., Ke, X., Li, H., & Liu, X. (2024). Highly sensitive zinc oxide nanoparticle composite film with deep learning-assisted mobile technology for enhanced food freshness monitoring. Food Bioscience, 62. https://doi.org/10.1016/j.fbio.2024.105541
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
