Title :
Multidimensional algebraic-integer encoding for high performance implementation of DCT and IDCT
Author :
Dimitrov, V. ; Jullien, G.A.
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Calgary, Alta., Canada
fDate :
4/3/2003 12:00:00 AM
Abstract :
A recently introduced single variable algebraic integer encoding scheme allows low complexity error-free computation of the DCT and IDCT. Introduced is the use of multidimensional algebraic-integer encoding of the transformation matrix of the DCT and IDCT which increases the sparseness of the encoding matrix and substantially reduces the dynamic range of the transformation coefficients. The computational savings are approximately 20% for typical image compression applications.
Keywords :
algebraic codes; discrete cosine transforms; image coding; image reconstruction; sparse matrices; Chen 8-point DCT algorithm; DCT; IDCT; computational savings; discrete cosine transform; encoding matrix sparseness; high performance implementation; image compression; low complexity error-free computation; multidimensional algebraic-integer encoding; reconstruction accuracy; transformation coefficient dynamic range; transformation matrix;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20030406