Narushin, Valeriy G., Volkova, Natalia A., Dzhagaev, Alan Yu., Griffin, Darren K., Romanov, Michael N, Zinovieva, Natalia A. (2025) Coupling artificial intelligence with proper mathematical algorithms to gain deeper insights into the biology of birds’ eggs. Animals, 15 (3). Article Number 292. ISSN 2076-2615. (doi:10.3390/ani15030292) (KAR id:108490)
PDF
Publisher pdf
Language: English |
|
Download this file (PDF/2MB) |
Preview |
Request a format suitable for use with assistive technology e.g. a screenreader | |
Official URL: https://doi.org/10.3390/ani15030292 |
Abstract
Simple Summary
Most birds’ eggs are products of consumer demand (for direct consumption or used to produce edible birds, mostly chickens). Many scientists are thus engaged in analyzing their shape to help improve quality, productivity and marketability, and these studies open up prospects for the use of artificial intelligence (AI). Deep learning (DL) is a form of AI concerned with computers taking data and extrapolating them to new ideas without direct human guidance. DL is based on the workings of animal brains, and specifically performs classification, analysis and further scholarly tasks by learning from existing datasets. We first consider the “state of the art” of DL in the poultry industry, including image recognition and applications for detecting cracks, egg content and freshness. We comment on how AI algorithms need to be properly trained and consider egg profile geometry. We revisit previous publications’ egg shape mathematics, commenting on the pros/cons of each. Examining weight, volume, surface area and air cell calculations, we consider how DL might be applied to these. The future value of DL is in egg sorting before incubation, storage/incubation and dimension calculation. Combining mathematical models with AI/DL means that we are on the threshold of many scientific discoveries, technological achievements and industrial successes.
Abstract
Avian eggs are products of consumer demand, with modern methodologies for their morphometric analysis used for improving quality, productivity and marketability. Such studies open up numerous prospects for the introduction of artificial intelligence (AI) and deep learning (DL). We first consider the state of the art of DL in the poultry industry, e.g., image recognition and applications for the detection of egg cracks, egg content and freshness. We comment on how algorithms need to be properly trained and ask what information can be gleaned from egg shape. Considering the geometry of egg profiles, we revisit the Preston–Biggins egg model, the Hügelschäffer’s model, universal egg models, principles of egg universalism and “The Main Axiom”, proposing a series of postulates to evaluate the legitimacy and practical application of various mathematical models. We stress that different models have pros and cons, and using them in combination may yield more useful results than individual use. We consider the classic egg shape index alongside other alternatives, drawing conclusions about the importance of indices in the context of applying DL going forward. Examining egg weight, volume, surface area and air cell calculations, we consider how DL might be applied, e.g., for egg storage. The value of DL in egg studies is in pre-incubation egg sorting, the optimization of storage periods and incubation regimes, and the index representation of dimensional characteristics. Each index can thus be combined to provide a synergy that is on the threshold of many scientific discoveries, technological achievements and industrial successes facilitated through AI and DL.
Item Type: | Article |
---|---|
DOI/Identification number: | 10.3390/ani15030292 |
Uncontrolled keywords: | deep learning (DL); artificial intelligence (AI); machine learning (ML); avian egg studies; egg shape models; egg mathematical indices; egg volume; egg surface area; prediction of egg inner parameters; non-destructive testing |
Subjects: |
Q Science > Q Science (General) > Q335 Artificial intelligence Q Science > QA Mathematics (inc Computing science) Q Science > QH Natural history > QH324.2 Computational biology Q Science > QL Zoology S Agriculture > SF Animal culture |
Divisions: |
Divisions > Division of Natural Sciences > Centre for Interdisciplinary Studies of Reproduction Divisions > Division of Natural Sciences > Biosciences |
Depositing User: | Mike Romanov |
Date Deposited: | 22 Jan 2025 16:22 UTC |
Last Modified: | 23 Jan 2025 22:35 UTC |
Resource URI: | https://kar.kent.ac.uk/id/eprint/108490 (The current URI for this page, for reference purposes) |
- Link to SensusAccess
- Export to:
- RefWorks
- EPrints3 XML
- BibTeX
- CSV
- Depositors only (login required):