Embedded Zeroblock Coding of images without lists

We present an efficient implementation of the Embedded Zeroblock Coding (EZBC) algorithm for image compression called Listless EZBC (LSEZBC). The LSEZBC implementation avoids the high computation and memory costs due to the use of link-lists in EZBC by using fixed-size state buffers to keep track of the set partitionings and coding process. Large groups of predictable coefficients are efficiently skipped during the coding process by using a recursive scanning method. The new implementation requires less memory and lower complexity than EZBC but has similar performance. Moreover, LSEZBC is extended for shape-adaptive coding. Superior performance is demonstrated when compared with several prominent shape-adaptive coders.