Year: 2023
Type: article
Abstract: Better health outcomes have been achieved as a result of improvements in the ability to identify lung cancer. areutilised frequently in the field of medicine to detect lung tumours at an early stage. Deep learning models such asUNET, Efficient-net, Resnet, VGG-16, and others have been used in a number of studies to improve the accuracywith which lung cancer may be detected. This paper offers an approach that combines UNET and Efficient-Netneural networks for the purpose of lung nodule segmentation and classification. The goal of the algorithm is toimprove the detection performance. In order to make use of the massive volume of CT scan pictures that do nothave any pathological diagnoses attached to them, an approach that is feature-extraction based and semi-supervisedis applied. A feature pyramid network (FPN) with the ResNet-50 model is used for feature extraction, and a neuralnetwork classifier is used for predicting unlabeled nodules. This setup allows for semi-supervised learning to takeplace. The skip-connections are the primary innovation that UNET brings to the table. These connections grant thedecoder access to the features that the encoder learnt at different scales, which in turn enables accurate localizationof lung nodules. An effective neural network for image classification is produced by Efficient-Net by employingscaling on all three dimensions of the network (depth, width, and resolution) in conjunction with a compoundcoefficient that scales all of the network's dimensions consistently. This work has been evaluated on the LIDC-IDRIdataset, which is available to the public, and it performs better than the majority of the existing approaches. Issuessuch as a high false-positive rate, small nodules, and a wide range of non-uniform longitudinal data are some of theproblems that the suggested algorithm intends to address. The findings of the experiments indicate that this modelhas a better level of accuracy than earlier research, coming in at 91.67% of the time.
Author: Dr. Tao Tian
Institution: Nanjing University
Language: English
Topic: AI in cancer detection
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7220585037
