A Survey on Artificial Intelligence-based Lung Tumor Segmentation and Classification

Abstract

Lung Tumor (LT) is difficult to detect, making it a particularly dangerous type of cancer. As a result, quick and precise nodule assessment is more crucial for patients of both sexes. LT can now be treated using a wide range of techniques and diagnostics. The earlier the LT is detected, the better the prognosis for the patient. Typically, a pathologist review is utilized to identify a tumor, but this method is time-consuming and error-prone. The automatic detection of the tumor would be extremely beneficial to pathologists. There has been a proliferation of ways for identifying LT with the emergence of Computed Tomography (CT) scans and x-rays in recent years. This study compares and contrasts various Artificial Intelligence (AI) techniques like machine learning (ML) and deep learning (DL) methods for identifying LT. A combination of image recognition and segmentation algorithms can be used to find LT nodules. This paper also includes the metrics used to validate the classification and segmentation technique. Moreover, an overview of imaging modalities and publicly available benchmark databases utilized in prior LT investigations are discussed. This information will be helpful to anyone working in the relevant field.