Conference Agenda
Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).
|
Daily Overview |
| Session | |
|
MED 2: Methods For Medical Applications 2 Location: De Carli Session Chair: Prof. Michele Calì, Electric Electronics and Computer Engineering Department, University of Catania Session Chair: Dr. Maria Antonella Messina, Kore University of Enna | |
    | |
Lightweight Deep Learning Tool for Melanoma Pre-Screening: A Cost-Sensitive Approach University of Palermo, Italy Skin cancer, especially melanoma, is a crucial public health issue where ear-ly diagnosis could drastically improve survival rates. This paper presents a comparative analysis of two lightweight convolutional neural networks, Effi-cientNetB0 and EfficientNetB3, optimized for mobile applications, to support primary dermatological screening. To address the class imbalance in the HAM10000 dataset, a cost-sensitive learning approach combined with dy-namic data augmentation was adopted, prioritizing a conservative balancing strategy. The models were then evaluated using a recalibrated decision threshold to maximize clinical sensitivity for malignant lesions. Analysis re-sults showed that the EfficientNetB3 architecture, combined with the downsampling adopted strategy, achieves better performance. The findings demonstrate the viability of deploying optimized, balanced AI models on computationally restricted devices for accessible triage. | |