Vol. 12, Special Issue 12 (2023)

Author(s):
Salam Jayachitra Devi, Jaya Bharati and NH Mohan

Abstract:
Image segmentation is a critical task in biological image analysis, providing essential information for various biomedical applications. Manual segmentation, though accurate, is time-consuming and impractical for big image data. In this study, we analyze an automated approach to enhance live and apoptotic cell segmentation using FIJI, a popular open-source image processing platform. This method leverages supervised learning algorithm based on pixel classification with different filters to achieve precise and efficient segmentation of cells in complex biological images. Experiments were conducted on different image data such as pig luteal cell microscopy image with trypan blue stain and without stain. Comparative analysis of the results obtained from different filters with manual segmentation is performed. From the analysis, the pixel-based classification algorithm performs better in case of image with trypan blue stain that achieved sensitivity of 0.93% for apoptotic cell and 0.86% for live cell. Average IoU score of pixels based segmented cells to manual segmented cells is above 0.91 for apoptotic cell and 0.80 for live cell. This paper contributes to the field by providing a comprehensive framework for automated cell segmentation in FIJI, paving the way for improved efficiency in biological image analysis. The method not only enhances segmentation accuracy but also showcases the potential for wider applicability in diverse biological imaging contexts.