Implementation of Efficient Segmentation Method for Histopathological Images

In biomedical research domain, it is always very crucial to examine the tissue affected by disease under a microscope and determine the severity of the disease with the process called histopathology, where the labs can produce tissue slides for viewing images digitally. These histopathological images has now become fully digitized. Digitizing these slides, allows pathologist to view them on a computer rather than viewing them on the microscope. Hence cell nuclei recognition and classification plays a crucial role in early detection of cancer. This task becomes even more challenging due to its heavy noise, and small-variant sizes of cell nuclei in histopathological images. To address this issue, an optimization based super pixel clustering algorithm was employed for automatic nuclei cell segmentation. Initially, the histopathological images dataset is acquired from suitable database. And then, normalization technique is applied to remove the noise from images. After denoising, segmentation is applied by using an optimized clustering algorithm to separate the non-nuclei and nuclei cells. The main aim of the proposed method is to implement an efficient segmentation method to overcome the issues of overlapping cells by Segmenting the histopathological images of lung/breast/liver/brain cancer, the proposed work has attained good performance to further help in the early detection of cancer by extracting the relevant features for classifying the given images as benign and malignant.