DIC images

Early-stage cervical cancer is characterized by morphological changes in cells, which are effectively identified through biopsy techniques with high diagnostic accuracy. However, biopsies can be expensive and occasionally painful, with diagnosis reports often taking days or weeks to generate. These delays and costs create significant barriers for underprivileged women with limited access to timely and affordable healthcare. Our Cervi- ImagingDiag framework and app provide a painless, cost-efficient, and accessible solution, delivering diagnostic reports within seconds. The app’s lightweight design ensures compatibility with limited-resource devices while safeguarding sensitive patient data and maintaining privacy. Our proposed framework and app integrate the lightweight federated CerviImagingYOLO architecture for cervical cancer prediction in the early stages, along with the CerviImagingLangChain framework to provide detailed diagnostic explanations. The lightweight federated CerviImagingYOLO model achieved an 81% accuracy rate, outperforming lightweight federated YOLO versions YOLOv5 (80%), YOLOv6 (76.47%), YOLOv8 (75.20%), YOLOv9 (80.00%), YOLOv10 (75.70%) and other lightweight federated architectures like MobileNetV3 (61.48%), SqueezeNet (71.66%), and EfficientNet (56.30%). The CerviImagingLangChain framework enhances patient understanding through detailed explanations. The app allows doctors to upload differential interference contrast images, review automated diagnoses, and share treatment plans, while patients receive diagnosis reports, schedule appointments, and access consultations.