MBC-based shape retrieval: basics, optimizations, and open problems

Shape of an object is an important feature for image and multimedia similarity retrievals. In our previous studies we introduced a new $$boundary$$-$$based$$ technique ($$MBC$$-$$based$$) for shape retrieval and compared its performance to other techniques. In this study, we describe in detail the basics of our $$MBC$$-$$based$$ shape representation techniques, and we show how they support different query types. In addition, we describe two original optimization techniques that can further improve the performance of our $$MBC$$-$$based$$ methods in several aspects, and show that they are also applicable to other applications (e.g., pattern recognition techniques). Finally, we define open problems in the area (e.g., partial similarity) and provide some hints on how to approach those problems.