Approaches to static background identification and removal

A common feature of many image processing applications is the classification of the data into regions of interest, often called the foreground, and unwanted regions referred to as being the background. An example of this process is the extraction of the steady-state component from a video sequence. Subtraction of this static component yields images containing only the desired foreground data. This is a common operation in real-time image processing performed within widely varying applications. As image processing develops, both in complexity and application, the role of background extraction algorithms in the pre-processing of raw data is increasing. The reasons for using these techniques can be categorised loosely into three key areas: bandwidth reduction, cost, and complexity reduction