How is the structure of an image that has to be coded organized?
First of all, JPEG2000 makes use of the DWT (discrete wavelet transform), whereas JPEG baseline is based on the use of the DCT (discrete cosine transform). The DWT transforms iteratively one signal into two or more filtered and cut signals that correspond to different frequency bands. Essentially, the image is divided into four spectral bands (LL, LH, HL, HH) that are obtained by the means of a low-pass (L) or high-pass (H) filtering along the horizontal and vertical directions. Afterwards, each of these four bands is divided once more into four other subbands; the operation is repeated a number of times that depends on the number of levels d of the DWT. The obtained samples are grouped into blocks (the code-blocks) and these blocks carry out the decodification of the bit-stream. Obviously, the larger the blocks size will be, the higher possibility of exploiting the redundancy among the samples (that is a higher coding efficiency) there will be. On the other hand, the use of blocks of
