“No more blurred blown-up images!” by Kim, Park and Lee

  • ©Hongseok Kim, Chang Joon Park, and In-Ho Lee




    No more blurred blown-up images!



    Popular interpolation methods introduce new pixel values which didn’t initially exist in the original image, and these new values across edges critically degrade overall perceptual image quality. Many approaches including [Li and Orchard 2001] and [Kimmel 1999] have tried to get crispier edges by giving more weights to pixels along edges, not across. However, the final enlarged result images almost always look blurred or water-colored paintings rather than photos, and it is quite obvious when we compare gradients maps of original and result images, which are quite different.


    1. Kimmel, R., Demosaicing: Image Reconstruction from CCD Samples, Proc. Trans. Image Processing, 1999, vol. 8, pp. 1221–1228.
    2. Li, X. and Orchard, M. T., New Edge-Directed Interpolation, IEEE Trans. Image Processing, 2001, vol. 10, no. 10

ACM Digital Library Publication:

Overview Page: