Loading...
Comparisons of pap smear classification with deep learning models
Author(s)
Date Issued
April 1, 2019
Type
Conference Paper
Abstract
We presented a comparative work of deep learning models for Pap smear classification. The benchmark parameters used to compare are accuracy, specificity, computation time, and sensitivity. Five convolution neural network models were employed to compare performance in detecting the presence of cervical precancerous or cancerous cells from a Pap smear database. The best deep learning model for multiclass classification is the densenet161 with an efficiency of 68.0% which will use to implement in our custom-made whole slide imager.
Citation
Proceedings of the 14th Annual IEEE International Conference on Nano Micro Engineered and Molecular Systems NEMS 2019, 282-285, 2019