Abstract

BRAVO-ORTIZ, MARIO ALEJANDRO et al. Cervical cancer classification using convolutional neural networks, transfer learning and data augmentation. Rev.EIA.Esc.Ing.Antioq [online]. 2021, vol.18, n.35, pp.100-111.  Epub Nov 03, 2021. ISSN 1794-1237.  https://doi.org/10.24050/reia.v18i35.1462.

Cervical cancer is formed in the cells that line the cervix and the lower part of uterus. Due to the cost and low reasons and low supply of services for the detection of this type of cancer many women do not have access to an early an accurate diagnosis. With the purpose of solving this issue ir was created a certain method that helps us to automatically classify the different types of cervical cancer, such as mild type 1 and 2, and aggressive (type 3), using digital image processing techniques and deep learning. We have a built a computational model based on convolutional neural networks, transfer learning and data increase, which help us obtain a classification accuracy up to 97.35% on the validation data, thus, we can ensure the reliability of the results. With this work it was demonstrated that the proposed design can be used as a complement to improve the tools of the assisted diagnosis of cancer.

Keywords : data augmentation; cervical cancer; convolutional neural networks; transfer learning.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )