應用於X光影像之雜訊移除與動態範圍強化處理

Abstract

放射線影像對於醫療病患的診斷佔有一個很重要的地位，而影像對比強化技術能夠讓X光影像有更好的對比度以及視覺的可視性以利於醫生能夠更準確的從放射線資料中診斷出病人的病情所在，然而，隨著影像的強化程度往往伴隨而來的是使原始影像產生影像的失真以及雜訊連帶的被放大，如此反而會造成醫療診斷上的誤差。因此存在著雜訊抑制的需求以及適當對比強化技術的必要，現今醫療影像強化論文大都是只針對單一雜訊來處理其雜訊抑制的問題，在本論文研究之主要目的在於提出一個兩段式濾波的影像強化技術應用於X光影像的對比動態範圍強化上面，使得就算X光影像存在著混合類型的雜訊，也能夠藉由適應性中值濾波器與雙邊濾波器的搭配完善的移除雜訊且良好的保留邊緣資訊，且利用數學形態學的頂帽、底帽轉換應用初步強化X光影像的細節對比度，接著利用對比限制適應性直方圖等化方法分塊修剪並重新分配局部直方圖像素點的分布來改變其累積密度函數再執行直方圖等化的方式來達到在整體光線亮度不失真且平坦區域微幅雜訊不被放大的情形下良好的達到理想的局部對比度與可視性。本論文演算法在針對多張混合雜訊的X光影像模擬測試上皆有良好的結果。

Radiological images plays a very important role in the medical diagnosis. The image enhancement allows X-ray images have a good contrast and visibility for medical examinations. It allows doctors to diagnose the condition of patient from the radiation data accurately. However, the degree of enhancement applied to the original image could produce distortions and amplify noise. That will lead to misinterpretation. Therefore, noise suppressions and an appropriate contrast enhancement techniques are necessary. This paper have proposed an algorithm of dynamic range enhancement with two stages filtering for medical images. Even X-ray images have mixed noise, our method can remove the noise completely and preserve the edge excellently by using adaptive median filter and bilateral filter. Then it enhances the contrast by using top-hat and bottom-hat transform and contrast limited histogram equalization methods. In simulation experiment, our algorithm has good results in several X-ray image with mixed noise.