Foveated Imaging

The human eye has a "high" resolution area at the center, and the rest of the eye is relatively low resolution. Can this biologically inspired technique be adapted from computer vision? An object of interest is saved at high resolution and the remaining scene is saved at low resolution. Once the object of interest is identified in the scene, it is deleted from the original model and only its coordinates are saved. Credit to Peter Gibbons for telling me about this. A bit of digging shows lots of papers including Wang et al. "Foveation scalable video coding with automatic fixation selection"

Category:

Science & Technology

Tags:

• computer
• vision
• image
• processing
• compression