|
KMID : 1191420110100020071
|
|
Korean Journal of Optometry and Vision Science
2011 Volume.10 No. 2 p.71 ~ p.77
|
|
|
Image Processing for Artificial Retina Using Saliency Map and Skin Detection
|
|
Chang Min-Hye
Kim Hyun-Seok
Jin Seon-Mi
Park Kwang-Suk
|
|
|
Abstract
|
|
|
Purpose: Because of the technical limitation, the number of microelectrode used for visual prosthesis is limited. Therefore we tried to deliver better visual information expected to attract viewer’s attention efficiently by applying a saliency map and skin detection algorithm.
Methods: Seven subjects who had normal sight participated. Their gaze movement was detected about each method. After using histogram equalization, saliency map was applied to original image. Skin detection algorithm was used only when there was human in the image.
Results: When using saliency map, subject’s gaze movement corresponded much more with attended region in original image than those without the methods by 48 percent. For image with a person, subject’s gaze movement corresponded more with attended region in original image by 6 percent.
Conclusions: Saliency map and skin detection delivered attended visual information more simply and more quickly.
|
|
|
KEYWORD
|
|
|
Artificial retina, Saliency map, Skin detection
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|
|