Selective visual attention based clutter metric with human visual system adaptability

摘要

Most existing clutter metrics are proposed based on fixed structural features and experienced weight measures. In this paper, we identify the clutter as selective visual attention effects and propose a type of clutter metric. First, adaptive structural features are extracted from the blocks with an edge-structure similarity to the target. Next, the confusing blocks are selected by the similarity threshold based on the attention guidance map. The clutter is estimated by quantifying the effects of confusing blocks on target acquisition performance. The comparative field experiments, with a Search_2 dataset, show that the proposed metric is consistent with the adaptability of the human visual system (HVS) and outperforms other metrics. (C) 2016 Optical Society of America