Segmentation Quality for Augmented Reality: An Objective Metric
MetadataShow full item record
Assessment the quality of segmentation algorithms considering the user perception is an important problem in Computer Vision. For this purpose a metric must take into account the impact of different types of errors displayed to the users. In this work we developed a new objective metric to assess the quality obtained by bilayer segmentation algorithms when they are used in Augmented Reality applications. Our metric proved to be efficient for selecting the best algorithm and to adjust its parameters.