This paper describes an algorithm to implement image filtering and enhancement operations by processing adaptive triangular meshes that represent gray-level images. Experimental results show that these operations are significantly more efficient when they are performed upon triangular meshes than by sequentially processing all the pixels contained in the given images.