標題: Segmentation of confocal microscopic image of insect brain
作者: Wu, MJ
Lin, CY
Ching, YT
資訊工程學系
Department of Computer Science
關鍵字: active contour;snake;gradient vector flow;segmentation;confocal microscopic image
公開日期: 2002
摘要: Accurate analysis of insect brain structures in digital confocal microscopic images is valuable and important to biology research needs. The first step is to segment meaningful structures from images. Active contour model, known as snakes, is widely used for segmentation of medical images. A new class of active contour model called gradient vector flow snake has been introduced in 1998 to overcome some critical problems encountered in the traditional snake. In this paper, we use gradient vector flow snake to segment the mushroom body and the central body from the confocal microscopic insect brain images. First, an edge map is created from images by some edge filters. Second, a gradient vector flow field is calculated from the edge map using a computational diffusion process. Finally, a traditional snake deformation process starts until it reaches a stable configuration. User interface is also provided here, allowing users to edit the snake during deformation process, if desired. Using the gradient vector flow snake as the main segmentation method and assist with user interface, we can properly segment the confocal microscopic insect brain image for most of the cases. The identified mushroom and central body can then be used as the preliminary results toward a 3-D reconstruction process for further biology researches.
URI: http://hdl.handle.net/11536/18779
http://dx.doi.org/10.1117/12.467125
ISBN: 0-8194-4429-4
ISSN: 0277-786X
DOI: 10.1117/12.467125
期刊: MEDICAL IMAGING 2002: IMAGE PROCESSING, VOL 1-3
Volume: 4684
起始頁: 1563
結束頁: 1570
顯示於類別:會議論文


文件中的檔案:

  1. 000177471900168.pdf

若為 zip 檔案,請下載檔案解壓縮後,用瀏覽器開啟資料夾中的 index.html 瀏覽全文。