An intelligent lung cancer detection from computed tomography images using robust optimal deep transfer learning with XAI model

Description

Lung cancer is found too late too often, and CT scans hold early signs that tired eyes can miss. Prabha, Poonkodi, and Premalatha assemble a deep learning pipeline that reads lung scans and explains its answers.

The pipeline first cuts the lung out of each scan with a segmentation model, then enriches the features and prunes them with an optimization routine modeled on shark smell. A transfer-learned network, InceptionResNetV2, makes the final call. To keep clinicians in the loop, an explanation module paints heat maps over the regions that drove each decision. On a public scan collection the system beat the comparison methods on accuracy, sensitivity, and speed. The authors pitch it as a reliable, interpretable aid for earlier diagnosis.

Catalog record

Source: https://europepmc.org/articles/PMC13352904