Scene Segmentation Using the Wisdom of Crowds

摘要: Given a collection of images static scene taken by many different people, we identify and segment interesting objects. To solve this problem, use the distribution in along with new field-of-view cue, which leverages observation that people tend to take photos frame an object interest within field view. Hence, image features appear together are likely be part same object. We evaluate effectiveness cue comparing segmentations computed our method against hand-labeled ones for several models. also show how results can used highlight important objects label them using noisy user-specified textual tag data. These methods demonstrated on popular tourist sites downloaded from Internet.