Does the scanning pattern affect adaptive quantization processing?

In this paper, we investigate whether different scanning pattern affect the image quality reconstructed from adaptive quantization processing. Different space filling curves, such as Peano curve, Hilbert curve, Moore curve, are implemented and compared in terms of compression ratio as well as image quality expressed in PSNR. The investigated space filling curves preserve the spacial neighborhood property of the pixel array, which is useful for predictive coding. While using those space filling curves, higher reconstructed image quality can be expected compared with the traditional raster scanning scheme.