Related Experiment Video
Updated: Jun 5, 2026
05:16
Flying Insect Detection and Classification with Inexpensive Sensors
Published on: October 15, 2014
25.1K
Automated identification of Chagas disease vectors using AlexNet pre-trained convolutional neural networks
Vinícius L Miranda1, João P S Oliveira-Correia2, Cleber Galvão2
1Laboratório de Parasitologia Médica e Biologia de Vetores, Faculdade de Medicina, Universidade de Brasília, Brasília, Brazil.
Medical and Veterinary Entomology
|December 13, 2024
Summary
Deep learning accurately identifies Chagas disease vectors (Triatominae subfamily). This automated system aids in controlling the spread of Trypanosoma cruzi, the parasite causing Chagas disease.
Area of Science:
- Entomology
- Medical Entomology
- Computer Science
Background:
- Triatominae subfamily insects are vectors of Trypanosoma cruzi, the cause of Chagas disease.
- Automated identification systems for these vectors require extensive image databases.
- Accurate vector identification is crucial for Chagas disease control.
Purpose of the Study:
- To evaluate the performance of a deep learning network (AlexNet) for identifying triatomine species using dorsal images.
- To assess the feasibility of developing an automated Chagas disease vector identification system.
Main Methods:
- A dataset of 6397 dorsal images of adult triatomines from seven genera and 65 species across 27 countries was used.
- The AlexNet deep learning model was trained and evaluated on this image dataset.
- Performance was measured by accuracy in identifying triatomine species.
Main Results:
- AlexNet achieved an overall accuracy of approximately 0.93 (95% CI, 0.91-0.94) for species identification.
- Highest accuracy was noted for species within the Rhodnius and Panstrongylus genera.
- Performance improved to approximately 0.95 (95% CI, 0.93-0.96) when focusing on species with the highest vectorial capacity.
Conclusions:
- Deep learning models like AlexNet demonstrate excellent performance for identifying triatomine species when trained on large, diverse image datasets.
- This study supports the development of an automated system for Chagas disease vector identification.
- The findings contribute to enhanced surveillance and control strategies for Chagas disease.
