腦神經結構重建視覺化編輯工具之研發

Abstract

神經的樹狀結構資料在神經科學及生物工程上有著重要的應用，在科學研究上有從影像的形狀或是結構來取得資訊做進一步分析的需求，但是從原始的影像資料要做許多形狀或是結構的分析是較為困難的，所以發展出了以全自動的方式取得影像的樹狀結構，但若是以全自動的方式來取得影像中的結構，還是有許多的因素將會影響到得到的結構品質，甚至有些影像會得到失敗的結果，所以要如何提升成果的精確度以及成功率是相當重要的。 本論文主要為實作一個影像結構重建的視覺化編輯系統，並加入一些可控制的因素，這些因素能夠提高影像在被追蹤出來之後的精確度以及成功率。本系統除了提供影像結構的追蹤和影像編輯的功能之外，還提供了一個讓使用者簡單驗證及比對結果的介面。

Neuronal tree structure on bio-engineering, neuroscience has important applications, In scientific research, from the image shape or structure to obtain information is very important, but analysts many shapes or structures from the original image data is more difficult. Therefore, the development of a fully automated way to obtain images of the tree structure, but the image structure we obtain from this way. There are many factors that affect the structure of the quality of the resulting. Therefore, how to improve the quality of the outcome is important. This study implements a visualization editing tool for image structure reconstruction, and add some controllable factor, that can promote the quality of result. The system is not only has the feature of tracing and editing, but also provide a simple validation interface.