Change detection is one of the most commonly used tasks in video processing. Background subtraction based change detection is the first step in many video applications to detect the foreground objects. Most of the background subtraction methods such as Local Binary patterns (LBP) and Local Ternary Patterns (LTP) are implemented using pixel by pixel representation. These can be viewed as improvements in many cases but they suffer in complexity, processing speed and illumination variations. In this paper, Local Binary Similarity Patterns (LBSP) makes pixel level decisions with automatic adjustment of tuning parameters for locally adapting to the changing input. It assigns the binary codes based on the similarity. For dynamic texture analysis, each pixel is modelled as a group of Spatiotemporal LBP (STLBP) histograms which combine spatial texture and temporal motion information together. The feedback loops are used to adjust the internal parameters and these adjustments are based on the continuous monitoring of model fidelity and segmentation noise which is observed under the form of blinking pixels. Adaptive background subtraction methods address all types of real time challenges including sudden illumination variations, background movements, shadows and ghost artifacts (falsely classified background regions).