Wing Interferential Patterns (WIPs) and machine learning, a step toward automatized tsetse (Glossina spp.)...

Area of Science:

• Veterinary Entomology
• Medical Entomology
• Bioinformatics

Background:

• Accurate identification of tsetse fly (Glossina spp.) species is essential for controlling Human African Trypanosomiasis (HAT) and African Animal Trypanosomiasis (AAT).
• Current identification methods rely on expert knowledge or destructive, costly molecular techniques, limiting widespread field application.
• There is a need for a simple, cost-effective, and accurate method for identifying Glossina species in endemic regions.

Purpose of the Study:

• To develop and validate a novel, cost-effective methodology for the accurate identification of Glossina species.
• To assess the feasibility of using Wing Interference Patterns (WIPs) combined with deep learning for automated tsetse fly classification.
• To provide a tool that supports the sustainable management of HAT and AAT.

Main Methods:

• A database of Wing Interference Patterns (WIPs) was created, comprising 1766 images of wings from 23 Glossina species.
• A deep learning architecture was employed to analyze the WIPs database for species classification.
• The methodology involved mounting wings on slides and utilizing a commercially available microscope.

Main Results:

• The developed deep learning model achieved very high accuracy in automatically recognizing Glossina species based on WIPs.
• The WIPs method proved to be a cost-effective alternative to existing identification techniques.
• The study successfully demonstrated the potential of WIPs as a reliable medium for automated Glossina species identification.

Conclusions:

• Wing Interference Patterns (WIPs) combined with deep learning offer a highly accurate and cost-effective solution for identifying Glossina species.
• This novel methodology can significantly aid in the field identification of tsetse flies, supporting HAT and AAT control programs.
• The approach presents a promising, non-destructive, and accessible tool for entomological surveys in Neglected Tropical Disease research.