M.C. Park (Korea)
Saliency map, region of interest, motion vector, viewer, and small Display
This paper presents a saliency-based smart image viewer for small displays. The suggested image viewer is very useful to browse some visually interest or stimulus regions to human visual system. It is hard to discern the resolution of an original image for small display because the image is shown in a state of spatial compression. The suggested method provides users with automatic image browsing ROI (Region of Interest) based on the saliency of an image. ROI is shown with the resolution of a display without any spatial compression. Therefore users are able to not only observe ROI automatically but also discern the original image quality. The regions of interest are shown in order according to the intensity of artificial visual attention. A saliency map model computes ROI with the intensity of the visual attention. To activate a typical saliency map model for moving images motion vector fields are added to a typical saliency map model while some algorithms are simplified to save computation time. A saliency-based image viewer is implemented for small displays like PDAs. The implemented image viewer is very useful to show a large image for small displays in that it shows ROI of an image according to the intensity of artificial visual attention.
Important Links:
Go Back
