Dynamic Sliding Window Width Selection Strategies for Rate-Distortion Optimal Vertex-Based Shape Coding Algorithms

Vertex-based operational-rate-distortion (ORD) shape coding algorithms frequently use a sliding window (SW) to both avoid trivial solutions and improve computational efficiency, with the choice of the SW-width enabling the encoder to trade between bit-rate and computational complexity. This paper presents some new strategies for dynamically determining the SW-width adaptive within the rate-distortion constraints. From a constrained bit-rate perspective, this is achieved by estimating the maximum number of significant points feasible for a prescribed bit-rate, while for an admissible distortion, an additional rate trade-off parameter selects the SW-width, with the optimal width determined for an encoder constrained by both rate and distortion. An efficient shape-adaptive technique is also presented that exploits the curvature of a shape to compute the SW-width. Experimental results confirm that the SW-width determined using the proposed strategies can be seamlessly and efficiently embedded into the ORD shape coding framework