具自動化指骨切割之基於Tanner and Whitehouse方法骨齡判讀雲端服務系統研製

Implementation of a TW3-Based Bone Age Assessment Cloud Service System with Automatic Phalanx Segmentation

楊又權(You-Cyuan Yang)

指導教授 ： 鐘太郎

國立清華大學/電機資訊學院/電機工程學系/碩士(2014年)

摘要

在醫療研究上，骨齡評估可以了解幼童生長情形。藉由骨齡不僅可以得知孩童是否有生長遲緩，甚至可以幫助兒科診斷內分泌系統。以往骨齡評估方法以Tanner and Whitehouse (TW)法和Greulich and Pyle (GP)法較常被使用。GP骨齡評分法需要專業醫師將待測手骨X光與資料庫內的圖像一一比對，找出最相似的圖片得到對應骨齡，其結果容易受到個人的主觀影響。而TW骨齡評估法是取左手X光片影像，再對左手X光影像內的二十個骨頭部位一一評分。一般而言，TW法較為客觀，但相當地耗費時間。雲端運算因其便利性在近年來相當地流行，因此本實驗室已有一套利用PHP和MySQL將TW骨齡評分法應用到雲端伺服器上的系統。但此系統在功能上還有不完備之處且在評分上還是需要花費相當多時間，所以本研究致力於改善系統不便之處及降低TW骨齡評分法所需的時間。若能將TW3法所需評分的骨頭部位先行取出，評分者就不必選取對應的手骨X光影像位置，達成簡化步驟的目的。已有研究可將手骨X光中的指骨部位影像取出，但在有手指併攏或重疊的情況下則無法正常表現。因此本研究提出不論五指在何種情形(正常、併攏或重疊)下，皆可將指骨影像正確取出的方法。有了適用於多數手骨X光影像的自動化指骨切割方法後，將其加入本TW3骨齡評分系統，可減去了評分者找尋及圈選待評估的位置的時間，使得TW骨齡評分更有效率。

關鍵字

骨齡評估 ； PHP ； MySQL ； Tanner and Whitehouse method ； 指骨切割

並列摘要

In the pediatric applications, bone age assessment is used to evaluate the growth of children. From the bone age, we can not only know that if there is growth retardation in the children, but also predict the adult height of the children. Tanner and Whitehouse III (TW3) method is a method of assessing skeletal age of children. In general, TW3 method has a benefit in reducing subjective factors. However, it will be quite time-consuming because there are twenty bone regions on left-hand X-ray image needed to examine against standard radiograms of each growth stage. In recent years, with the development of computer network, some web-based TW3 scoring systems could be found in the Internet. Although it reduced use of papers when scoring bone age by using TW3 method, evaluators or physicians still need to spend a lot of time scoring their data.

In order to simplify the procedure of TW3 method, it is beneficial to develop a pre-processing algorithm to partition the hand radiograms into regions. Previous studies have reported how to separate the phalanx on radiograms from the background and soft tissue. However, if the fingers on the radiograms overlapped, they could not do the job rightly. In this study, we propose a phalanx segmentation algorithm to extract phalanx ROI fast and precisely. Furthermore, when dealing with the case of radiograms with fingers overlapped, our algorithm works as good as the usual. Moreover, the cloud computing becomes popular by its large data storage property. Thus, a bone age assessment cloud computing system, which is based on our phalanx segmentation algorithm and TW3 method, was built. We hope the physicians or radiologists can evaluate the bone age by TW3 method more efficient with the help of this system.

並列關鍵字

bone age assessment ； Phalanx segmentation

參考文獻

[1] J. M. Tanner, “Assessment of skeletal maturity and prediction of adult height (TW2 method).” London; New York: Academic Press, 1975.

[2] W. W. Greulich and S. I. Pyle, “Radiographic atlas of skeletal development of the hand and wrist.” Stanford, CA: Stanford University Press, 1959.

[3] W. W. Greulich and S. I. Pyle, “Radiographic Atlas of Skeletal Development of the Hand and Wrist,” 2nd edition. Stanford, CA: Stanford University Press, 1959.

[4] J. M. Tanner, H. Goldstein, N. Cameron, W. B. Saunders, and M. J. R. Healy, “Assessment of skeletal maturity and prediction of adult height (TW3 method)”, 3rd ed. W.B. Saunders, 2001.

[6] D. G. King, D. M. Steventon, M. P. O'Sullivan, A. M. Cook, V. P. Hornsby, I. G. Jefferson, and P. King, “Reproducibility of bone ages when performed by radiology registrars: an audit of Tanner and Whitehouse II versus Greulich and Pyle methods, “ British Journal of Radiology, vol. 67, Sept. 1994.