Segmented images and surface models made of serially-sectioned images of cadaver head

PURPOSE: Purpose of this research is to present serially-sectioned images of cadaver head with 0.1 mm intervals and 0.1 mm pixel size, accompanied by coronal and sagittal images, which are helpful in interpreting MRIs. Another purpose is to present segmented images and surface models (three dimensional images) based on the serially-sectioned images, which are helpful in learning head anatomy. The other purpose is to present technique of erasing horizontal lines in coronal and sagittal images, and that of segmentation and surface reconstruction with automation extent incresed, which help other researchers make their own medical images. MATERIALS & METHODS: Horizontal serially-sectioned images (0.1 mm intervals, 0.1 mm pixel size, 48 bit color) of cadaver head were made by serial sectioning cadaver head, into which fixative or dye was not injected, and then reformed. Coronal and sagittal images were acquired by stacking the serially-sectioned images, and then horizontal lines in the vertical images were automatically erased. Anatomical structures including detailed brain structures were labeled in serially-sectioned images, coronal images, and sagittal images. 7 Tesla MRIs of a living person were scanned, which were similar with the serially-sectioned images of cadaver head. Anatomical structures were also labeled in horizontal, coronal, and sagittal planes of the MRIs. Segmentation technique, the automation extent of which was increased by quick selection tool and interpolation, was developed. By using the technique, 64 head structures were outlined to make temporary segmented images (PSD file, 1 mm intervals) and segmented images (TIFF file, 1 mm intervals). Volume models of the segmented structures were created by volume reconstruction on free software. Advanced surface reconstruction technique was developed, the automation extent of which was enhanced via volume reconstruction. By using the technique, surface models of the segmented structures were created. Segmented images, in which outlines of the structures were combined, were made and then included in browsing software of the serially-sectioned images. RESULTS: A total of 2,343 horizontal serially-sectioned images (0.1 mm intervals, 0.1 mm pixel size, 48 bit color) of the cadaver head were prepared. Coronal and sagittal images without horizontal lines were prepared. 7 Tesla MRIs of living person’s head were prepared, which were similar with the serially-sectioned images. Horizontal, coronal, and sagittal planes of the seriallysectioned images and MRIs with head structures labeled were prepared. 234 couples of temporary segmented images and segmented images (1 mm intervals) of 64 head structures were prepared. Volume models of the structures were prepared. 64 surface models of the structures were prepared. Browsing software, which showed serially-sectioned images and corresponding segmented images, were prepared. CONCLUSION: The serially-sectioned images of cadaver head as well as the labeled horizontal, coronal, and sagittal images are helpful in interpreting MRIs of brain. The segmented images, volume models, and surface models are helpful in learning head anatomy. The images will encourage other researchers to make various software, such as virtual dissection or virtual operation, which are expected to contribute to medical education. The browsing software are helpful in learning sectional anatomy. Technique of erasing horizontal lines of coronal and sagittal images, and that of enhancing automation of segmentation and surface reconstruction, developed in this research, will help other researchers make their own medical images.

Author dc:creator

• 신, 동선

Contributors dc:contributor

• 정, 민석
• 대학원 의학과
• 105819

• 연속절단면영상
• 시신 머리
• 구역화영상
• 부피3차원영상
• 표면3차원영상

Language dc:language

ko

Identifier

http://dcoll.ajou.ac.kr:9080/dcollection/jsp/common/DcLoOrgPer.jsp?sItemId=000000009326
000000009326

OAI identifier oai:identifier

oai:repository.ajou.ac.kr:201003/1825