ADVANCING LUNG CANCER PREDICTION:INNOVATIVE AI MODULES

Computer Engineering

Lung cancer is the leading cause of cancer deaths worldwide. Directly diagnosing the malice of suspected lung nodes is of consummate clinical significance. Accurate discovery of lung nodes on casket reckoned tomography reviews is pivotal to early opinion of lung cancer. To address the nuisance problems on low discovery perceptivity and high false-positive rate caused by diversity and morphological complexity of 2- D bump features, a computer- backed discovery system is developed to increase the discovery perceptivity and bracket delicacy of lung nodes. Aiming at the disadvantages of low delicacy and slow speed in current lung bump discovery styles, this design proposes a featherlight lung bump vaticination algorithm grounded on bettered Region Convolutional Neural Network. This system first preprocesses the original CT image and sends it to the Region Offer Network for pulmonary nodes discovery. also, the discovery results are used to complete the bracket of benign and nasty lung nodes through the Region grounded Convolutional Neural Network (R- CNN). The experimental results on the LUNA16 data set prove that the bettered network model chart value can reach96.71, and the discovery speed can reach41.99 FPS. Results interlace that our proposed bracket network could give an effective individual tool for suspected lung nodes and might have a promising operation in clinical practice.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.