Image Classification Via LZ78 Based String Kernel: A Comparative Study

Normalized Information Distance (NID) [1] is a general-purpose similarity metric based on the concept of Kolmogorov Complexity. We have developed this notion into a valid kernel distance, called LZ78-based string kernel [2] and have shown that it can be used effectively for a variety of 1D sequence classification tasks [3]. In this paper, we further demonstrate its applicability on 2D images. We report experiments with our technique on two real datasets: (i) a collection of real-life photographs and (ii) a collection of medical diagnostic images from Magnetic Resonance (MR) data. The classification results are compared with those of the original similarity metric (i.e. NID) and several conventional classification algorithms. In all cases, the proposed kernel approach demonstrates better or equivalent performance when compared with other candidate methods but with lower computational overhead.