Related Experiment Video
Updated: Apr 11, 2026
06:57
Loop-mediated Isothermal Amplification LAMP Assays for the Species-specific Detection of Eimeria that Infect Chickens
Published on: February 20, 2015
27.7K
Classification of chicken Eimeria species through deep transfer learning models: A comparative study on model
Zeki Kucukkara1, Ilker Ali Ozkan1, Sakir Tasdemir1
1Selcuk University, Faculty of Technology, Department of Computer Engineering, Konya, Türkiye.
Veterinary Parasitology
|January 24, 2025
Summary
Deep transfer learning (DTL) models accurately classify Eimeria species in chickens. The Xception model achieved 96.4% accuracy, offering a potential automated diagnostic tool for coccidiosis control.
Area of Science:
- Veterinary Parasitology
- Computational Biology
- Animal Science
Background:
- Eimeria causes coccidiosis in chickens, leading to significant economic losses.
- Conventional Eimeria species identification is challenging due to similar oocyst morphology and time-consuming methods.
- Accurate species identification is crucial for epidemiological studies and effective disease control.
Purpose of the Study:
- To develop an automated system for classifying digital images of sporulated Eimeria oocysts.
- To utilize deep transfer learning (DTL) models for accurate Eimeria species identification in chickens.
- To evaluate the performance of various DTL models for this classification task.
Main Methods:
- Utilized 17 pre-trained deep transfer learning (DTL) models.
- Applied feature extraction and fine-tuning methods for image classification.
- Classified digital micrographic images of sporulated Eimeria oocysts from seven pathogenic chicken species.
Main Results:
- The Xception DTL model achieved the highest classification accuracy at 96.4%.
- DTL models demonstrated significant potential for classifying Eimeria species.
- The developed system outperformed other tested models.
Conclusions:
- Deep transfer learning models, particularly Xception, are highly effective for automated Eimeria species classification.
- This approach can reduce researcher workload and aid in developing diagnostic tools for coccidiosis.
- The DTL models show promise for practical applications in parasitology and other scientific fields.
