Adaptive fast spatial averaging filter

Spatial averaging filter is one of the most extensively used techniques in image processing at different levels. Various algorithms for spatial averaging are proposed to improve its performance. These are primarily focused on reducing the number of additions required for spatial averaging computation. Performance of these algorithms is found inconsistent for different sizes of images even though total number of pixels is same. We proposed a method, called Adaptive Fast Spatial Averaging (AFSA) that employs an adaptive sliding window mechanism based on width and height of an image. AFSA is applied on various images with different filter sizes and showed that its performance is consistent to total number of pixels. Proposed method shows that its performance is consistent to total number of pixels irrespective of the dimensions of an image.