A study on improved deep learning structure based on DenseNet

초록

The existing image-related deep learning research methods are conducted through algorithms based on feature identification and association, but there are limits to their accuracy and reliability. These methods are inefficient for artificial neural networks to extract features and learn because of the loss of spatial information in the process of removing background and flattening images and have a limit on increasing accuracy and reliability. The deep learning algorithm applied in this study was based on the DenseNet neural network which is recently the best in performance and accuracy, and its architecture was improved with a focus on increasing the learning performance. As a result of the experiment, both speed and accuracy of learning data were more increased than the existing DenseNet architecture, which means to diagnose more images than the existing methods within the same amount of time. Copyright © 2022, IGI Global.

키워드