An Objective Scoring Framework for Histology Slide Image Mosaics Applicable for the Reliable Benchmarking of Image Quality Assessment Algorithms
| Autor: | Roxana M. Buga, Natasa Sladoje, Tiberiu Totu, Adrian Dumitru, Stefan G. Stanciu, Mariana Costache |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
0301 basic medicine
medicine.medical_specialty Microscope General Computer Science Computer science Image quality image quality assessment ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION image stitching Rendering (computer graphics) law.invention Image stitching 03 medical and health sciences law Microscopy image mosaics medicine Digital pathology General Materials Science Bright-field microscopy General Engineering Histology Benchmarking 030104 developmental biology Histopathology lcsh:Electrical engineering. Electronics. Nuclear engineering lcsh:TK1-9971 Algorithm |
| Zdroj: | IEEE Access, Vol 6, Pp 53080-53091 (2018) |
| ISSN: | 2169-3536 |
| DOI: | 10.1109/access.2018.2868127 |
| Popis: | The conversion of histology slides into electronic format represents a key element in modern histopathology workflows. The most common way of converting physical histology slides into digital versions consists of tile-based scanning. In such approaches, the entire image of the slide is generated by consecutively scanning adjacent sample regions with a degree of overlap and then stitching these together to constitute an image mosaic. To achieve a high-quality result, the image acquisition protocol for collecting the mosaic tiles requires a recalibration of the microscope when moving from one sample region to another. This recalibration procedure typically involves focus and illumination adjustments, aimed at rendering a homogeneous image mosaic in terms of brightness, contrast, and other important image properties. The accurate evaluation of the digital slide's quality factor is, therefore, an important matter, as it can lead to designing efficient (and automated) mosaic generation protocols. We introduce here a new methodology for the evaluation of image mosaics collected with brightfield microscopy on histology slides, coined Objective Quantifiable Scoring System (OQSS). It relies on objective scoring criteria that take into consideration fundamental characteristics of image mosaics, and on histology specific aspects. We present the theoretical principles of this methodology and discuss the potential utility of this framework as a quality ground-truth tagging mechanism of histology slide image mosaics applicable for the reliable benchmarking of image quality assessment algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|