Department of Analytical Chemistry , National Institute of Chemistry , Hajdrihova 19 , 1000 Ljubljana , Slovenia.
Anal Chem. 2018 May 1;90(9):5916-5922. doi: 10.1021/acs.analchem.8b00751. Epub 2018 Apr 9.
This work focuses on the structural similarity (SSIM) index as a tool for optimization of the perceived visual image quality obtainable by continuous scanning 2D LA-ICPMS bioimaging, but also other mass spec imaging techniques may benefit from this approach. This index quantifies the differences between a distorted image and a reference image based on parameters associated with luminance, contrast, and noise. Since reference images are not normally available, a protocol was developed to virtually apply distortion-related information introduced by the LA-ICPMS imaging system to a reference image of one's choice. Distortion-related information in the form of blur and noise was experimentally retrieved from line scans across a laser milled knife edge on custom-prepared gelatin standards (mimicking proteinaceous biomatrixes). Distorted images were generated via computational procedures developed earlier, warranting objective image quality assessment via the SSIM indices. We illustrate the potential of this approach for image quality optimization for a suite of LA-ICPMS imaging conditions.
这项工作专注于结构相似性（SSIM）指数，作为通过连续扫描二维 LA-ICPMS 生物成像获得可感知视觉图像质量的优化工具，但其他质谱成像技术也可能受益于这种方法。该指数根据与亮度、对比度和噪声相关的参数来量化失真图像和参考图像之间的差异。由于通常无法获得参考图像，因此开发了一种协议，以便将由 LA-ICPMS 成像系统引入的与失真相关的信息虚拟应用于所选的参考图像。以模糊和噪声形式的失真相关信息是从在定制制备的明胶标准（模拟蛋白质生物基质）上的激光铣削刀刃上的线扫描中实验获得的。通过早期开发的计算程序生成失真图像，通过 SSIM 指数保证客观的图像质量评估。我们说明了这种方法在一系列 LA-ICPMS 成像条件下进行图像质量优化的潜力。
