CASIA OpenIR  > 毕业生  > 博士学位论文
扩散光学断层成像研究及其应用
其他题名Progress and Application of Diffuse Optical Tomography
赵青
学位类型工学博士
导师蒋田仔
2007-04-26
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业模式识别与智能系统
关键词扩散光学断层成像 多中心光纤排布模式 Lsa算法 基于反s型的光子密度调整 深度分辨率 大脑成像 乳腺成像 Diffuse Optical Tomography Multicentered Mode Layer-based Sigmoid Adjustment Depth Resolution Brain Imaging Breast Imaging Image Quality
摘要扩散光学断层成像技术是一种无创的人体组织功能检测手段。它可以达到毫秒级的时间分辨率,设备成本低廉,然而,该技术还有成像区域较小、深度分辨率较低等缺点。本文的目的正是解决扩散光学断层成像技术的这些缺点,推动该技术的临床实用化。 近两年来,多中心光纤排布模式的出现大大改善了扩散光学断层成像技术的空间分辨率,然而,六边形的多中心光纤排布模式覆盖区域较小,要真正进行大脑成像,必须覆盖更大的区域。本文在六边形多中心光纤排布模式的基础上提出了五边形多中心光纤排布模式,并将五边形和六边形通过适当的方式拼接,形成覆盖区域更大的全脑成像头盔。 扩散光学断层成像技术的另一个缺点是深度分辨率较差,这主要是因为光子在人体组织中传播时,能量迅速衰减,导致深层的光子密度远小于浅层组织的光子密度。本文提出了一种基于反S型的光子密度调整算法(LSA算法),它通过在深层的光子密度上乘一较大的系数,在浅层的光子密度上乘一较小的系数来减小深层和浅层的光子密度对比度,从而提高扩散光学成像技术的深度分辨率。 本文将LSA算法应用在大脑成像中,结果表明该算法可以显著提高图像的深度分辨率,将大脑激活区定位在大脑皮层上。在应用了LSA算法的重建图像中,我们可以分辨出手指运动与对指任务激活区的差别,但在没有采用LSA算法的重建图像中,这两个任务激活区的区别不明显。 扩散光学断层成像技术还可以用在乳腺成像中,但是较低的深度分辨率导致重建图像中的噪声较大,图像质量不高。本文的结果表明,在应用了LSA算法之后,重建图像中的噪声显著减小,图像质量大幅提高。应用了LSA算法之后,扩散光学断层成像技术在检测较小的肿瘤和更早期的肿瘤方面的表现得到显著提升,有利于该技术的临床应用。
其他摘要Diffuse optical tomography (DOT) is an effective tool for imaging biological tissues. Comparing to other imaging methods, DOT has better temporal resolution and lower instrumental cost. However, DOT has some disadvantages such as limited imaging region and poor depth resolution. This paper focuses on overcoming the disadvantages of DOT. This will benefit its clinical applications. In recent two years, the multicentered mode has been proposed for improving the spatial resolution of DOT. However, the hexagon shaped multicentered mode has limited imaging region. To image adult human brain, we must expand the imaging region of the multicentered mode. In this paper, we proposed a pentagon shaped multicentered mode. Furthermore, we connected hexagons and pentagons to form a full-brain headgear. The poor depth resolution is another shortcoming of DOT. This is mainly because that photons lose there energies sharply as they propagate longer distance. In this paper, we proposed a layer-based sigmoid adjustment (LSA) method. The LSA method multiplies larger coefficients on deep tissues and multiplies smaller coefficients on superficial tissues, so the photon sensitivity contrast can be effectively reduced. In this paper, we implemented the LSA method in brain imaging. The results from simulations and human subject experiments showed that the depth resolution of the reconstructed images had been improved. In images reconstructed by LSA method, the activations are properly located on cerebral cortex and different activations can be observed for finger-moving and finger-tapping experiments. However, in images reconstructed without the LSA method, activations are biased to superficial layers and difference in activations can be hardly seen. Another important field DOT can be used in is breast imaging. In breast imaging, the poor depth resolution of DOT can cause large noise in reconstructed images. The results from simulations showed that by using the LSA method, noises in reconstructed images could be effectively reduced. With the LSA method, DOT can detect smaller and earlier tumor. This will benefit the clinical application of DOT.
馆藏号XWLW1106
其他标识符200418014628018
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/5962
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
赵青. 扩散光学断层成像研究及其应用[D]. 中国科学院自动化研究所. 中国科学院研究生院,2007.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20041801462801(3041KB) 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[赵青]的文章
百度学术
百度学术中相似的文章
[赵青]的文章
必应学术
必应学术中相似的文章
[赵青]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。