主要综述高分辨血管成像技术及其在生物医学领域中的应用，侧重评述适用于高分辨血管图像的定量表征方法。血管图像定量表征主要包括3个步骤：图像预处理、血管图像重建及定量特征获取、定量参数的统计学分析。同时，对每个步骤中所涉及的算法流程、准确性评估及后续算法的优化方向进行详尽的阐述。此外，探讨多种血管和血流参数所反映的生物学信息在临床上的参考意义，并结合具体的疾病场景，介绍多参数分析模型在区分不同疾病发展阶段方面的能力。本文的阐述不仅体现了高分辨血管成像技术及定量表征方法的潜在价值，也展示了它们在推进生物医学基础研究和临床诊断等方面的光明前景。

提出一种光栅 X 射线相衬断层扫描中吸收信号环形伪影的去除方法。该方法采用正弦图域和重建图域结合的处理算法。对于弱伪影, 通过正弦图域的排序和滤波去除。而对于强伪影, 首先根据环形伪影在极坐标系的表现, 计算残差图像, 转换到笛卡尔坐标系得到伪影像素和样品的边界; 进而使用基于机器学习的图像分割方法获取每一类样品的分布, 同时为了保护边界信息, 通过形态学操作获得样本的内部区域; 最后再利用残差图像的分布特征定位伪影像素, 并使用临近非伪影像素均值替代。实验结果表明该方法可以在不破坏样品边界的前提下有效地去除图像中的环形伪影。

同步辐射红外(SRIR)光具有光谱范围宽、发散角小、亮度高以及信噪比高等优点,结合传统红外谱学技术,采用SRIR谱学显微技术对样品进行红外谱学显微,可以获得样品微米级别的空间光谱信息。利用MiTeGen聚亚酰胺小环作为样品,以上海光源BL01B1线站的SRIR光为光源,通过点扫描采样方式进行同步辐射红外三维谱学显微实验研究。通过获得聚亚酰胺小环在不同角度下的SRIR二维显微光谱信息,选取波数范围为1495~1485 cm -1的显微光谱信息处理,用代数迭代算法对聚亚酰胺小环的化学组分酰胺Ⅱ进行SRIR三维显微重构,获得了完整的三维重构图。实验表明本文方法能够以较高的信噪比重构出样品化学组分的三维红外显微结构。

In this Letter, a single wavelength digital holographicmethod is proposed to achieve depth resolved imaging by recording a series of holograms in the reflection geometry with an illumination of constantly changing curvature. A proper algorithm is employed to selectively generate the images of the object at different depths, including the phase and the modulus information. Theoretical analysis is supported by a visible light experiment to illustrate the feasibility of the proposed method.

We outline the use of digital holographic tomography to determine the three-dimensional (3D) shapes of falling and static objects, such as lenslets and water droplets. Reconstruction of digitally recorded inline holograms is performed using multiplicative and Radon transform techniques to reveal the exact 3D shapes of the objects.

In landmark-based image registration, estimating the landmark correspondence plays an important role. In this letter, a novel landmark correspondence estimation technique using mean shift algorithm is proposed. Image corner points are detected as landmarks and mean shift iterations are adopted to find the most probable corresponding point positions in two images. Mutual information between intensity of two local regions is computed to eliminate mis-matching points. Multi-level estimation (MLE) technique is proposed to improve the stability of corresponding estimation. Experiments show that the precision in location of correspondence landmarks is exact. The proposed technique is shown to be feasible and rapid in the experiments of various mono-modal medical images.

Content-based image retrieval has been an active area of research for more than ten years. Gabor schemes and support vector machine (SVM) method have been proven effective in image representation and classification. In this paper, we propose a retrieval scheme based on Gabor filters and SVMs for hepatic computed tomography (CT) images query. In our experiments, a batch of hepatic CT images containing several types of CT findings are used for the retrieval test. Precision comparison between our scheme and existing methods is presented.

The operational procedures for efficiently reconstructing the two-dimensional image of a body by the filtered back projection are described in this paper. The projections are interpolated for four times of original projection by zero-padding the original projection in frequency-domain and then inverse fast Fourier transform (FFT) is taken to improve accuracy. Nearest interpolation is applied to decrease operation time. Projection dependence of next pixel at the same row is used. For each row of image, the first pixel projection once for each angle is pre-computed, other pixel projection of this row can be found by iteration. Therefore, the pre-interpolation process only has to be performed once for each row, rather than individually for each pixel. It greatly reduces the amount of computation. Compared with original implementation, the speed of reconstruction image is nearly improved by five times. The image accuracy is still preserved.

A retrieval scheme for liver computerize tomography (CT) images based on Gabor texture is presented. For each hepatic CT image, we manually delineate abnormal regions within liver area. Then, a continuous Gabor transform is utilized to analyze the texture of the pathology bearing region and extract the corresponding feature vectors. For a given sample image, we compare its feature vector with those of other images. Similar images with the highest rank are retrieved. In experiments, 45 liver CT images are collected, and the effectiveness of Gabor texture for content based retrieval is verified.