Due to a limited bit-depth representation of pixel values and lossy compression methods, an image may exhibit annoying banding or false-contouring artifact. Existing banding-removal approaches mainly include low pass filtering, dithering and requantization. In this paper, we propose an image debanding method that iteratively improves the image using a contentadaptive spatial filter. The filter uses adaptive span based on local width of the band and operates on 1D sparse pixel-window. The input image is filtered in each direction i.e. vertical, horizontal and two diagonals sequentially while offering parallel processing in each direction. The iterative framework gradually improves the quality of filtered image until it converges to the final output. Experimental results show that our proposed method is effective in debanding images.