Towards Resource-Certified Software: A Formal Cost Model for Time and its Application to an Image-Processing Example

Visual tracking requires sophisticated algorithms working in real-time, and often space-limited, settings. While the input streams may be regular in structure, the algorithms are not, and must often deal with probabilistic metrics. To ensure progress in algorithm design without incurring excessive development costs, we propose a high-level programming approach married with predictable and compositional performance metrics. This enables the combination of independently developed program components into coherent software architecture, with certified resource use guarantee. Here, we present our approach and discuss its application to the development and resource analysis of a space bound mean shift algorithm for motion tracking, using the new embedded system-oriented language Hume.