AI-DRIVEN APPLICATION FOR BREAST CANCER DETECTION

ABSTRACT Breast cancer is a situation whereby abnormal breast cells grow uncontrollably or in a different way that isn't normal, forming tumors. The purpose of the topic is to develop an AI-DRIVEN APPLICATION FOR BREAST CANCER DETECTION using deep learning algorithms. To achieve this, the system used yolov8 model to detect the breast cancer and show where the tumor is found or where its located. YoloV8 was trained using MRI dataset, nextjs,typescript and tailwind was used to design the user interface (UI), bounding box was utilized to locate or single out where the tumor was located. In the detection page the user have to upload the MRI image after which axios sends a post request to yolov8 API, API performs the detection and gives responses, if there's no error in the detection it will display the result of the detection using bounding box and then the user can choose to download the result, and if there was an error in the detection, the system will return an error handler which will take the user back to upload another MRI image again. Along the line, the performance evaluation matrix was used to evaluate the model training which gave the result such as precision of 1.0, F1 of 0.94 and recall of 0.96. This study achieved the development of an AI driven application for breast cancer detection and met the aims and objectives which was listed out in the initial chapter of this system.