Real-time cervical cancer risk assessment via mobile colposcopy and AI integration

Cervical cancer is one of the most common and dangerous cancers in women, especially in countries with limited resources. In this context, the principal aim of this project is to develop a tool that allows clinicians to screen for cervical cancer through the screening and processing of colposcopy images captured with a smartphone-based colposcope, Cervix app. The mobile application for cervical cancer diagnosis was developed using React Native and Firebase, enabling compatibility with iOS and Android devices. The application features a user-friendly and intuitive interface that facilitates the capture and analysis of colposcopy images. The classification and processing of images (benign and malignant) were conducted using the UNET model for segmentation, GANs for data augmentation, and ResNet models for classification. Several tests were conducted to evaluate the performance of the mobile application to predict and diagnose, ensuring its functionality was accurate and reliable at 90%.