The present invention relates to a magnetic resonance knee cartilage measurement method combining ICP and KD-tree registration algorithm for knee cartilage measurement. First, building a 3D image from the knee magnetic resonance sequence images obtained from MRI scanning, and subjecting the 3D image to an optimization process. Then morphology is used to capture the tibia surface of knee, and registration process of the 3D image is carried out with ICP and KD-tree open source code, followed by using a cellular automaton to segment the 3D image that has completed registration, and finally a fusion process is performed. The method of this invention can effectively solve the problem of segmentation failure and excessively long operation time caused by high similarity between knee cartilage pixels and the neighboring tissue pixels, and can assist a doctor in diagnosis and treatment determination.本發明是一種結合ICP及KD-TREE對位演算法於膝蓋軟骨量測的磁共振膝蓋軟骨測量方法,首先將MRI掃描所產出的膝蓋磁共振序列影像組成三維圖像,且予以影像最佳化處理,再運用型態學原理抓取膝蓋的脛骨面,然後利用ICP及K-D TREE開放式原始碼進行三維圖像的對位處理,再將對位完成後的三維圖像進行細胞自動機分割,最後進行融合之動作,而本發明方法可有效解決膝蓋軟骨像素與其周遭組織像素太過於相近所導致的分割失敗及操作時間過長等問題,且有利於於醫師診斷治療判斷。(a)、(b)、(c)、(d)、(e)‧‧‧步驟流程
