Improved Particle Swarm Optimization for Detection of Pancreatic Tumor using Split and Merge Algorithm

• Published in: Computer methods in biomechanics and biomedical engineering. Vol. 10; no. 1; pp. 38 - 47
• Main Authors: Dhruv, Bhawna, Mittal, Neetu, Modi, Megha
• Format: Journal Article
• Language: English
• Published: Taylor & Francis 02.01.2022
• Subjects: computed tomography; image segmentation; Pancreatic tumour; particle Swarm optimisation; split-merge
• ISSN: 2168-1163; 2168-1171
• DOI: 10.1080/21681163.2021.1966650

Pancreatic cancer is the fourth leading cause of cancer-related death worldwide. Pancreatic tumours are characterised by a peculiar cell progression in intestinal enzymes and hormone-producing cells. Patients with pancreatic tumours have a shorter survival period depending on the type of tumour they have. Enhancing patient survival, diagnosing patients with malignant pancreatic tumours as soon as possible is the need of hour. The commonly used approach for diagnosis and clinical disease staging of pancreatic cancer is Computed Tomography (CT). However, because most abdominal CT images contain noise in addition to visceral fat in the vicinity of the pancreas, prompt detection is difficult, which requires an algorithm that distinguishes the anatomical structure clearly. This paper proposes, Particle Swarm Optimisation incorporating Split and Merge Segmentation to obtain enhanced CT images of pancreatic tumour. The image undergoes image processing procedures such as segmentation using PSO and fetch an enhanced output image with the help of gaussian filter. The algorithm outperforms Conventional PSO in terms of Entropy and Contrast, according to experimental results. The improved CT scans aid in the early detection and diagnosis of pancreatic cancer.