Mesh-free technique for enhancement of the lung CT image

Lung cancer is abnormal cell disease and a leading cause of mortality for both men and women according to Global cancer statistics 2020 (GLOBACON). Pulmonary nodules appear to be one of the first signs of lung cancer which needs to be recognized for early diagnosis. CT images are distorted due to various noises and poor resolution. Image pre-processing plays an important role to get rid of noises and improve image quality. Enhancement of lung CT image facilitates to determine the sign of early stage of lung cancer in terms of tumor size, malignancy, rounded hematoma. A partial differential equation based heat diffusion equation is used in this work for enhancement of lung CT-scan images. Mesh-less based radial basis function technique is used to solve the heat equation. Proposed work employed on randomly selected sample images of two lung dataset named as The Cancer Genome Atlas Lung Adenocarcinoma(TCGA-LUAD) and QIN. The performance of the proposed lung enhancement technique is quantified by various parameters namely, Structure Similarity Index Measure (SSIM), Peak Signal-to-Noise Ratio (PSNR), and Contract Improvement Index (CII). The observational investigations demonstrated that the proposed technique is computationally more efficient as compared to FDM, Median, and CLAHE.