In this article, we present a method called adaptive patch adjustment that
improves the performance of digital image data compression using triangular
plane patches. The method is designed to adaptively adjust the three-dimen
sion position of triangular plane patches that approximate the correspondin
g luminance curved surfaces of an original image. Such an adjustment consid
ers the influence of all pixels contained in the projection of a patch, ins
tead of defining a patch by only three pixels at the vertices. The influenc
e is described and calculated with a soft computing technique, the simplifi
ed fuzzy reasoning. Experimental results show that such a flexible strategy
significantly reduces the average distortion between the reconstructed ima
ge and the original one, avoiding excess block splitting, and as a result,
increasing the data compression rate. (C) 1999 Elsevier Science B.V. All ri
ghts reserved.