一种基于显著图的针对625通道人造视觉的刺激模式图片提取.pdfVIP

Vol. 4 No.7/ Jul. 2010 一种基于显著图的针对625通道人造视 觉的刺激模式图片提取算法 A Stimulus Pattern Extraction Algorithm Based on Saliency Map for A 625-Channel Retinal Prosthesis System 陈丽 邹月娴 郑亚莉 摘  要 为了恢复盲人的部分视力，数千电极的视觉假体已经研制成功。视觉假体的图像处理系统将从相机获得的原始图片转换 到刺激模式，通过刺激芯片刺激大脑神经从而产生幻觉。实际上，从相机获得的图片是高分辨率的（256x256），而刺激模式图 片目前分辨率很低（25x25），这种从高分辨率到低分辨率刺激模式图片的提取是一个很大的技术挑战。本文着重于研究一种有 效的基于单通道显著图的刺激模式提取的算法，并优化了剪裁阈值的选择。实验结果表明，我们提出的刺激模式提取算法对于不 同场景都有非常好的性能。最后我们给出对于不用场景的最佳剪切阈值的选择建议。 关键词 人造视觉；图像处理；感兴趣区域；显著图；剪裁的阈值选择 ABSTRACT　In an effort to restore visual perception in retinal diseases, an electronic retinal prosthesis with thousands of electrodes has been developed. The image processing of retinal prosthesis system converts the original images from the camera to the stimulus pattern that can be properly interpreted by the brain. Practically, the original images are with much high resolution (256x256) than that of the stimulus pattern (such as 25x25), which causes a technical challenge to extract the stimulus pattern from the original image. In this paper, we focus on developing an efficient stimulus pattern extraction algorithm by using the single cue saliency map for extracting salient objects in the image with an optimal trimming threshold. Experimental results show that the proposed stimulus pattern extraction algorithm performs quite well for different scenes in terms of the perception of the stimulus pattern. Some suggestions are also given on trimming threshold selection for different scenes. KEYWORDS Retinal Prosthesis; Image Processing; Region Of Interest; Saliency Map; Trimming Threshold Selection 1  引言 几种难以治愈的眼疾导致每年10万左右的人失 的像素只有几百。因而摄像头接收的高分辨率的图像 明[1] 。两种最常见的外层视网膜退化疾病是年龄相 必须缩小为较低的分辨率(只有几百像素)而尽可能多的 关性黄斑变性(AMD)和视网膜色素变性(RP)。AMD 保留其中显著的物体/人。在这个转换过程中，