Desenvolvimento de técnicas utilizando CNN para detecção de câncer de pele

Abstract

Cancer is one of the leading causes of morbidity and mortality worldwide, posing an ongoing challenge to medicine and science. Given the increasing incidence of various types of cancer, enhancing diagnostic tools is crucial for improving survival rates and treatment effectiveness. This study proposes an approach that leverages advanced image processing techniques and deep learning for the segmentation and classification of different types of cancer. The methodology focuses on the application of Convolutional Neural Networks (CNNs), which are well known for their high performance in medical image segmentation and classification.Early and accurate detection is essential for the success of cancer treatment, yet traditional methods based on manual analysis face issues of subjectivity and inter-observer variability. To overcome these limitations, the study was structured into four main stages. Initially, the dataset was analyzed and prepared, involving the exploration of image characteristics and the application of techniques to balance class distribution. Next, computational models for lesion segmentation and classification were developed and fine-tuned, with an emphasis on improving diagnostic accuracy. Subsequently, the necessary infrastructure was implemented to integrate the models into an accessible and secure support system. Finally, a user-friendly interface was developed to allow image capture and interaction with the system, enhancing the user experience.The expected outcomes include the creation of an effective tool that can be integrated into clinical protocols, improving both the accuracy and speed of cancer diagnosis. By adopting these technologies, it is hoped not only to enhance diagnostic efficiency but also to positively influence patient treatment and prognosis. Thus, this study contributes significantly to the advancement of imaging-based diagnostic practices in oncology, exploring new technological frontiers to address one of the greatest challenges in public health.