Big data driven decision making and multi-prior models collaboration for media restoration

Abstract

Aiming at the restoration of degraded social network services media, this paper proposed a novel multi-prior models collaboration framework for image restoration with big data. Different from the traditional non-reference media restoration strategies, a big reference image set is adopted to provide the references and predictions of different popular prior models and accordingly guide the further prior collaboration. With these cues, the collaboration of multi-prior models is mathematically formulated as a ridge regression problem in this paper. Due to the computation complexity of dealing big reference data, scatter-matrix-based KRR is proposed which can achieve high accuracy and low complexity in big data related decision making task. Specifically, an iterative pursuit is proposed to obtain further refined and robust estimation. Five popular prior methods are applied to evaluate the effectiveness of the proposed multi-prior models collaboration. Compared with the traditional restoration strategies, the proposed framework improves the restoration performance significantly and provides a reasonable method for the relative exploration of big data driven decision making.