發(fā)布時間：2018-07-23 12:19

【摘要】：目的探討一種基于形態(tài)學(xué)重建和閾值分割運算的SPECT腎臟圖像感興趣區(qū)(ROI)自動勾畫方法。方法首先對SPECT腎動態(tài)顯像圖像進行預(yù)處理,提升圖像的對比度,再對其進行形態(tài)學(xué)重構(gòu),接著用閾值分割算法識別出腎臟,最后通過邊界跟蹤實現(xiàn)腎臟ROI區(qū)的自動勾畫。結(jié)果使用該圖像處理及分割算法實現(xiàn)的ROI自動勾畫,相比人工輔助勾畫,具有更高的準(zhǔn)確性。結(jié)論本文運用的形態(tài)學(xué)重建法和最大類間方差法,有較好的魯棒性,可以更準(zhǔn)確地計算腎小球濾過率,改善了現(xiàn)有Matlab腎動態(tài)定量分析軟件中ROI勾畫不準(zhǔn)確的情況,有助于進一步定量分析。
[Abstract]:Objective to study an automatic (ROI) mapping method for SPECT kidney image based on morphological reconstruction and threshold segmentation. Methods the SPECT renal dynamic imaging image was preprocessed to enhance the contrast of the image and then reconstructed by morphology. Then the kidney was identified by threshold segmentation algorithm. Finally, the ROI region of the kidney was automatically delineated by boundary tracking. Results the ROI automatic sketch realized by this image processing and segmentation algorithm is more accurate than that of artificial aided sketch. Conclusion the morphological reconstruction method and the maximum inter-class variance method used in this paper have better robustness, can more accurately calculate the glomerular filtration rate, and improve the inaccurate ROI description in the existing Matlab renal dynamic quantitative analysis software. It is helpful for further quantitative analysis.
【作者單位】： 南京醫(yī)科大學(xué)附屬南京醫(yī)院醫(yī)療設(shè)備處;
【基金】：南京市醫(yī)學(xué)發(fā)展基金“青年工程”人才培養(yǎng)專項經(jīng)費資助項目(QRX11033)
【分類號】：R817.4;TP391.41
，

本文編號：2139397