Special issue on recent advances in saliency models, applications and evaluations

The research on computational saliency models has been an active topic in the past decades, due to its broad applications in the fields of multimedia processing, computer vision, cognitive science, robotics, etc. Saliency models mainly target two genres of applications, i.e., human fixation prediction or salient object detection/segmentation. Although recent years have witnessed significant progress on saliency modeling for images/videos and the related applications, multimedia with new modalities such as RGB-D, stereoscopic, large-scale internet images/videos calls for tailored and effective saliency models, while images/videos with complex scenes need saliency models with higher performance. It is desirable to explore the use of saliency in a variety of interesting applications for improving their performances, and thus to effectively broaden the scope of saliency-based applications. Besides, new datasets, metrics, benchmarks and behavioral studies are expected to comprehensively evaluate and fairly compare the performance of state-of-the-art saliency models, as well as to shape the future research directions. Modeling the way we perceive multimedia content is fundamental for offering the best technologies to the emerging multimedia applications. Bearing this objective in mind, we invite original research articles to promote the research on saliency from models, applications and evaluations.