低視能者彩色視覺影像輔助系統之研發與分析

Abstract

視覺是人體中敏感的感覺器官之一，低視能者由於視覺上的限制，會造成生活機能與品質的降低。本研究為低視能患者提供一數位視覺影像輔助系統，藉由影像處理方式強化影像以增加患者對影像的辨識能力。影像強化方法以彩色影像中的明亮度色座標為主要強化影像，並將影像之塊狀與邊界特徵做強化，並將演算法延伸至同步彩色視訊影像處理以及同步AVI影片影像處理。接著為了驗證影像增強後對視覺補償的效果，本研究先以請十位受測者佩戴低視能模擬眼鏡並判斷所見圖形以及影片的方式進行驗證，接著再直接實際對十位弱視患者進行實驗。實驗結果顯示，我們可依低視能者視力退化程度對應影像強化參數，低視能者視力之退化程度是以科學化方式來分析視覺系統作了解，並藉由量測對比敏感函數(CSF)來得知視力系統衰退程度，並利用數值分析歸納出此強化演算法確實對低視能者有實質的影像資訊輔助，而結論告訴我們，大部分套用此輔助系統之影像對弱視患者都有相當程度的清晰度提升，但也會有少部分的人認為強化後的圖片有點失真的問題，細部特徵太過突兀，這將會是之後可繼續探討之問題。

Human vision is one of the sensitive sensory organ, therefore, the visual constraints will reduced the quality of life for low vision patients. The goal of this thesis is to provide a digital system enhancement for low vision, and we use image processing methods to increase the ability for patients to recognize the image. The method of algorithm which we enhance the bar and edge feature, which represent the medium and the high information, and we also mediate the parameters of feature component according to the measurement of contrast sensitivity function(CSF) of low visions. Besides, we use the algorithm to finish a real time color camera and AVI video processing system. In order to verify the effect of image enhancement for visual compensation, we invite ten people to wear low vision simulation glasses and ten low vision old people. The result showed that subjects have significant improvement to determine enhanced images, it shows that this algorithm indeed to achieve effective information quality improvement for low vision patients.