Related Experiment Video
Updated: Jun 16, 2026
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
1.5K
Automated Identification of Cutaneous Leishmaniasis Lesions Using Deep-Learning-Based Artificial Intelligence
José Fabrício de Carvalho Leal1,2, Daniel Holanda Barroso3, Natália Santos Trindade2
1Graduate Program in Tropical Medicine, Center for Tropical Medicine, Faculty of Medicine, University of Brasília-UnB, Brasília 70904-970, Brazil.
Biomedicines
|January 26, 2024
Summary
Deep learning (DL) using AlexNet accurately identifies cutaneous leishmaniasis (CL) skin lesions. This automated tool shows potential for mobile applications to aid healthcare professionals in diagnosing CL.
Area of Science:
- Dermatology
- Artificial Intelligence
- Medical Imaging
Background:
- Cutaneous leishmaniasis (CL) diagnosis is challenging due to varied lesion presentations, often mimicking other skin conditions.
- Accurate differentiation is crucial for timely and effective treatment.
Purpose of the Study:
- To evaluate the efficacy of AlexNet, a deep learning algorithm, in identifying images of CL skin lesions.
- To assess the potential of automated diagnosis for assisting healthcare services.
Main Methods:
- A dataset of 2458 CL lesion images from Midwest Brazil patients was compiled.
- The AlexNet model was trained and tested on this dataset, differentiating CL from 26 other dermatoses.
- The image database was split into training (80%), validation (10%), and testing (10%) sets.
Main Results:
- AlexNet achieved an average accuracy of 95.04% (95% CI: 93.81-96.04) in identifying CL lesions.
- The algorithm demonstrated excellent performance in distinguishing CL from various other skin conditions.
- Three simulations confirmed the model's robust identification capabilities.
Conclusions:
- Automated identification of CL using AlexNet shows significant potential to support clinical diagnosis.
- This technology could facilitate the development of mobile diagnostic tools for CL in healthcare settings.
- Further development could improve accessibility and efficiency of CL diagnosis.
