Challenges in developing a multidimensional Feature Selective Validation implementation

Feature Selective Validation (FSV) was incorporated as core component of IEEE Std 1597.1. Recently, discussions have moved from ‘is quantitative comparison of data needed?’ to ‘how can this be applied to multidimensional data?’ The purpose of this paper is to present some of the latest thinking about the key challenges in developing a 1D model to unlimited dimensionality. In particular, the paper will present: (1) a revised mathematical framework for FSV, e.g. using tensor notation to simplify the mathematical representation. (2) A discussion on how the performance of n Dimensional FSV can be verified against human perception. (3) A review on the effects of data that oscillates between positive and negative data points, e.g. transients.