A New Artificial Intelligence Approach for Early Skin Cancer Diagnosis

An international team of researchers, including experts from Politecnico di Torino, GPI SpA in Trento, University of Cadiz, University of Verona, COMSATS University Islamabad, and 7HC SRL in Rome, has developed an innovative system based on Artificial Intelligence (AI) to help specialists detect skin cancer at an early stage. This approach combines advanced algorithms and methods to achieve highly accurate results while reducing costs and time compared to traditional examinations.

Why It Matters

Skin cancer is one of the most common types of cancer worldwide, and early diagnosis is crucial. Identifying a lesion in its early stages can make the difference between effective treatment and a late-stage intervention. Until now, diagnosis has relied heavily on the trained eye of dermatologists, but this process can be time-consuming, costly, and somewhat subjective. This is where the new AI-powered system comes in, providing a fast and reliable second opinion.

How It Works

• Image Analysis: The program “reads” photos of skin lesions, identifying key characteristics for diagnosis.
• Optimized Algorithms: A deep learning model (complex neural networks) has been made lighter and faster so that it can run on less powerful computers.
• Result Explanation: Unlike other “black box” AI systems, this one allows users to see which areas the AI focused on, providing clear support for doctors and patients.

Key Results

• High Accuracy: The system correctly identifies skin cancer in about 98% of cases on one of the analyzed datasets.
• Reduced Time and Cost: Thanks to fast processing, this approach can lower the number of repeated visits and screenings.
• Greater Trust: With visual and textual explanations, dermatologists and patients can better understand how the diagnosis was made.

This research opens up exciting possibilities for early skin cancer detection using AI technologies. The goal is to support more dermatologists and specialists, improving healthcare quality and increasing the chances of early intervention before it’s too late.

Read the scientific article:
https://www.mdpi.com/2313-433X/10/12/332