A Deep Learning Approach for Classifying Developmental Stages of Ixodes ricinus Ticks on Images Captured Using a...

Area of Science:

• Entomology
• Computer Science
• Public Health

Background:

• Ixodes ricinus is Europe's most common tick and a significant vector for pathogens like Lyme disease bacteria and TBEV.
• Different tick developmental stages (larvae, nymphs, adult females, adult males) have varying roles in disease transmission.
• Nymphs are particularly epidemiologically relevant due to their prevalence and small size, making accurate identification crucial.

Purpose of the Study:

• To develop and validate a deep learning model for classifying Ixodes ricinus tick developmental stages.
• To address the need for automated identification of European tick species stages, differentiating from existing North American-focused solutions.

Main Methods:

• A convolutional neural network (CNN) model was developed and trained on microscopic images of Ixodes ricinus ticks.
• Image data was collected from the Upper Silesia region of Poland.
• Explainable AI (XAI) technique, Grad-CAM, was employed to visualize model decision-making processes.

Main Results:

• The CNN model demonstrated effectiveness in classifying the developmental stages of Ixodes ricinus ticks.
• The study is the first to apply CNNs for identifying European tick fauna developmental stages.
• Grad-CAM analysis provided insights into the image features critical for the model's classification accuracy.

Conclusions:

• The developed deep learning approach offers a novel and accurate method for identifying Ixodes ricinus tick developmental stages.
• This technology has significant potential applications in entomological research, public health surveillance, and tick-borne disease management.
• The model provides a valuable tool for distinguishing between tick stages, improving epidemiological assessments and control strategies.