Image segmentation is a common task which arises in many situations such as extracting a face or a character in a text from an image before performing automatic recognition. Generally, it is used to separate the foreground pixels belonging to the object(s) of interest, from the background pixels. Segmentation aims to partition a given image into a set of regions having following properties: connectivity and homogeneity in terms of color or texture.

An object or region can be a characterized using multiple low-level features such as brightness, colour, texture, etc. In this experiment, we will restrict ourselves to objects defined by brightness. We will aim to partition a given image into two (foreground and background) and study two techniques namely, histogram-based thresholding and region growing, to perform this partitioning.