FAST: Fuzzy atlas-based lung cancer segmentation technique from noisy medical images

Lung Cancer segmentation from noisy images is a complex task due to variations in feature extraction and pixel representations. However, identifying anatomical structures of cancer from such complex images is predominant in segmenting infected regions for diagnosis. This article introduces FAST, which emphasizes the concept of fuzzy atlas-based segmentation, a technique employed in image analysis, with a particular focus on medical imaging. This method integrates existing anatomical models, known as atlases, with fuzzy logic to enhance the precision of identifying and distinguishing various structures within noisy medical images. The anatomical structures of the noise-reduced and unreduced images are used to derive multiple levels of fuzzy derivatives to identify a precise infected pixel distribution. The primary advantages include the incorporation of established anatomical knowledge, the ability to accommodate individual differences, and the potential for improved outcomes in noisy images. This technique is particularly beneficial in medical imaging scenarios where accurate identification of anatomical structures is essential. The proposed technique’s efficacy is validated and proved using accuracy, precision, recall, and F-Measure metrics.