A Review of Object Segmentation Quality Subjective Assessment Methods

Ran Shi,King Ngi Ngan,Yichi Zhang
DOI: https://doi.org/10.1109/tencon.2015.7372726
2015-01-01
Abstract:In the past decades, researchers have made great efforts to develop object segmentation algorithms and constantly improving the segmentation quality. Compared with object segmentation algorithms, object segmentation quality assessment is less studied. Especially, how to evaluate segmentation quality by subjective assessment, which is the most reliable way to understand and analyze segmentation quality in terms of human perception has not attracted much attention. In this paper, we review three object segmentation subjective assessment methods and propose possible future directions for research in object segmentation quality subjective assessment.
What problem does this paper attempt to address?