Cervical Cell Nuclei Segmentation On Pap Smear Images Using Deep Learning Technique
S Alisha, Vinitha Panicker J · 2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT) · 2022
The fourth most common type of cancer in women is cervical cancer. The cells lining the cervix, the lowest portion of the uterus, are where cervical cancer first develops. Cancer detection that is accurate and timely can save lives. The Pap smear experiment is a form of cervical screening that looks for probably malignant and cervical or colon cancer precursors. Automated and precise segmentation of the cervical nucleus is required as nuclei contain crucial diagnostic details for automated cervical cancer monitoring and computer detection techniques. The deep learning model used for image segmentation in this research is the EfficientDet. By undertaking a rigorous analysis of earlier modern detection prototypes, it uncovered methods to enhance computing efficiency, which led to the creation of EfficientDet. We evaluated our model using the Herlev Pap Smear data set and observed that the precision and recall values obtained were above 0.9, and the mean IOU score was 0.836.