Computer vision in the analysis of lip cancer: Increasing diagnostic reliability with the use of Explainable Artificial Intelligence.

• Grant number: 24/13315-4
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: December 01, 2024
• Status: Discontinued
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: André Carlos Ponce de Leon Ferreira de Carvalho
• Grantee: Ayrton da Costa Ganem Filho
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated scholarship(s): 25/06148-7 - Agnostic Computer Vision-Based explainability applied to medical images., BE.EP.IC

Abstract

The use of Artificial Intelligence (AI), such as Convolutional Neural Networks (CNNs), has demonstrated efficiency and high assertiveness for medical applications. Despite this success, there are significant challenges to be overcome, related to the availability of "real-world data" and CNN architectures, which face challenges in terms of reliability and explainability, given that the use of AI in the health area requires interpretable results, thus providing confidence to health professionals and transparency that enables more accurate diagnoses. This project aims to explore the use of CNNs in the visualization of clinical images of lip cancer for diagnostic purposes, with a particular focus on squamous cell carcinoma of the lip (SCCL). The research emphasizes the generation of explainability estimates to support clinical decision-making, which generates confidence in the medical diagnosis. This approach allows for a deeper interaction between automated diagnosis and medical expertise, facilitating the creation of more effective protocols for cancer treatment. By exploring the intersection between advanced technology and medical practices, this study contributes to the evolution of cancer diagnosis, highlighting the importance of AI-based methods in modernizing public health.