Proposal and Evaluation of a Deep Learning Model for image segmentation

Abstract

Breast cancer represents a major global health challenge, particularly due to its impact on women's health. Improving diagnostic capabilities is essential to ensure timely intervention and reduce complications. Yet, the traditional manual analysis of mammographic images often proves to be time-intensive and susceptible to variability among medical professionals. This thesis explores a deep learning-based approach for the semantic segmentation of mammographic images, aiming to enhance the detection and localization of suspicious breast lesions. In addition to developing a robust segmentation technique, this work includes the integration of the system into a mobile user interface, enabling practical testing and accessibility. The proposed approach illustrates the growing role of artificial intelligence in advancing medical imaging and supporting early breast cancer diagnosis