The study of disease indicators through microscopic analysis of a biopsy or surgical specimen that has been prepared and preserved on glass slides is known as histopathology. In histopathology studies, nucleus segmentation is a pivotal task facilitating the analysis of nuclei morphology, classification of cell types, and critical functions like cancer detection and grading. The tissue sections are stained with one or more stains so that distinct tissue components can be seen under a microscope. One of the main tissue stains used in histology is hematoxylin and eosin (also known as hematoxylin-eosin stain; frequently abbreviated to H&E stain or HE stain). Manual nuclei segmentation faces various challenges because of the image background’s complexity, shape change, and nucleus occlusion or overlap. Deep learning-based models for automated nucleus segmentation have gained popularity due to their better segmentation task performance. In this article, a U-Net based model is used to collect multi-scale nuclei features and acquire context information about nuclei. U-Net models, introduced specifically for biomedical image segmentation, excel in its versatility in accepting different types of input data, such as grayscale, color, and multi-channel images. Variants of U-Net including U-Net++, U-Net3+, and Double U-Net are experimented on the dataset to compare their performances.