斑点图案
图像分辨率
像素
计算机视觉
计算机科学
时间分辨率
血流
人工智能
图像处理
光学
信号(编程语言)
各向异性
物理
图像(数学)
程序设计语言
内科学
医学
作者
Abhishek Rege,Janaka Senarathna,N. Li,Nitish V. Thakor
标识
DOI:10.1109/tbme.2012.2183675
摘要
Laser speckle contrast imaging (LSCI) is a full field optical imaging technique, capable of imaging blood flow without the introduction of any exogenous dyes. Spatial and temporal resolution in LSCI images depend on how pixels are chosen from the raw image stack for contrast processing. However, all processing schemes are based on isotropic treatment of the spatial neighborhood about each pixel, restricting further improvement in spatiotemporal resolution and image quality. We present a novel spatiotemporal processing scheme for LSCI where the spatial neighborhood is anisotropic, that is, restricted along a specific direction that matches direction of blood flow. The technique allows for a significant increase in temporal resolution, from conventionally used 40 or 80 frames to just three frames; while simultaneously achieving 23% and 47% higher signal-to-noise ratios over concurrent spatiotemporal schemes, when imaging rapid and slow functional changes in blood flow, respectively. We present the concept, justification, and performance evaluation of the novel scheme and demonstrate its suitability for imaging rapid changes in blood flow. Anisotropic LSCI was able to monitor the heart rate associated fluctuations in intravascular blood flow and showed them to be as high as 28% of the mean.
科研通智能强力驱动
Strongly Powered by AbleSci AI