Identification of Chicken Eimeria Species using Deep Learning Approaches

Boufenar, Chaouki; Djamai, Samir; Mezghiche, Dihia; Soudani, Lina

doi:10.1109/i3cis56626.2022.10075855

Cited by 2 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite the imbalanced distribution of sample numbers among species in the datasets, our preliminary study revealed that this imbalance did not negatively impact the classification results. Consequently, similar to other studies [9,31,33,[35][36][37] in the literature, we utilized the datasets without employing any sampling methods to balance them.…”

Section: Resultsmentioning

confidence: 99%

“…Not updating the parameters in the convolution layer in the fine-tuning of the models has also contributed to this goal. It is observed in the literature [9,34,37] that there is a common approach that has been taken to create two separate CNN models for the detection of rabbit and fowl parasites, separately. As this is mainly a parasite detection problem linked to microscopic images, a holistic approach has been taken to develop one single CNN model with the output of 17 classes and a similar number of CNN parameters in a single model.…”

Section: Discussionmentioning

confidence: 99%

“…[36,37] incorporated CNN models to classify the fowl species, Boufenar et al compared five different pretrained CNN models for the same purpose and found the Xception as the most proficient model. [9] Our study adopts a holistic approach to identify all varieties of Eimeria species leading to disease in both fowls and rabbits, diverging from the conventional practice of detecting them within distinct groups. Furthermore, this study not only determines the most effective pretrained model out of 27 diverse models within this domain but also makes it energy-efficient by implementing our pruning techniques.…”

Section: Classification Of Eimeria Parasitesmentioning

confidence: 99%

“…On the other hand, this study specifically focused on classifying Eimeria parasites, leading to serious diseases in many animals, especially rabbits and fowls. [9] Our approach involved leveraging pretrained CNN models to identify these parasites, which have many species. In this context, we created a framework employing 27 different pretrained models widely used in contemporary research and industry.…”

mentioning

confidence: 99%

See 3 more Smart Citations

Green AI‐Driven Concept for the Development of Cost‐Effective and Energy‐Efficient Deep Learning Method: Application in the Detection ofEimeriaParasites as a Case Study

Acmali,

Ortakci,

Seker

2024

Advanced Intelligent Systems

View full text Add to dashboard Cite

Although large‐scale pretrained convolutinal neural networks (CNN) models have shown impressive transfer learning capabilities, they come with drawbacks such as high energy consumption and computational cost due to their potential redundant parameters. This study presents an innovative weight‐level pruning technique that mitigates the challenges of overparameterization, and subsequently minimizes the electricity usage of such large deep learning models. The method focuses on removing redundant parameters while upholding model accuracy. This methodology is applied to classify Eimeria species parasites from fowls and rabbits. By leveraging a set of 27 pretrained CNN models with a number of parameters between 3.0M and 118.5M, the framework has identified a 4.8M‐parameter model with the highest accuracy for both animals. The model is then subjected to a systematic pruning process, resulting in an 8% reduction in parameters and a 421M reduction in floating point operations while maintaining the same classification accuracy for both fowls and rabbits. Furthermore, unlike the existing literature where two separate models are created for rabbits and fowls, this article presents a combined model with 17 classes. This approach has resulted in a CNN model with nearly 50% reduced parameter size while retaining the same accuracy of over 90%.

show abstract