Diagnostic quality assessment of medical images: Challenges and trends

Abstract

International audienceWith medical imaging technologies growth, the question of their assessment on the impact and benefit on patient care is rising. Development and design of those medical imaging technologies should take into account the concept of image quality as it might impact the ability of practicians while they are using image information. Towards that goal, one should consider several human factors involved in image analysis and interpretation, e.g. image perception issues, decision process, image analysis pipeline (detection, localization, characterization...). While many efforts have been dedicated to objectively assess the value of imaging system in terms of ideal decision process, new trends have recently emerged to deal with human observer perfomances. This task effort is huge considering the variability of imaging acquisition methods and the possible pathologies. This paper proposes a survey of some key issues and results associated to this effort. We first outline the wide range of medical images with their own specific features. Next, we review the main methodologies to evaluate diagnostic quality of medical images from subjective assessment including ROC analysis, and diagnostic criteria quality analysis, to objective assessment including metrics based on the HVS, and model observers. At last, we present another evaluation method: eye-tracking studies to gain basic understanding of the visual search and decision-making process

Medical image and video quality assessment in e-health applications and services

“…HVS models consider various factors like contrast sensitivity, luminance, masking effect, and frequency decomposition. The metrics usually correlate well with human visual performance and hence enhance the quality evaluation efficiency [32]. Also, an interesting observation from the reviewed works is that the traditional metrics like PSNR, SSIM do not perform well in the context of medical images [33].…”

Comprehensive Classification Model for Diagnosing Multiple Disease Condition from Chest X-Ray

“…Various research papers have discussed challenges associated with diagnosing medical images, but very few are found to be associated with chest x-ray [5] [6]. Over the period of time, certain researchers have dedicated themselves to discuss the existing challenges as well as problems associated with image processing techniques over chest radiographs [7][8] [9]; however, there is no effective disclosure about effectivity in the approaches presented by researchers. The automated diagnosis processes (on the basis of research) are categorized into two problems, i.e., identification process and classifying the disease.…”

High Dynamic Range Image Compression by Optimizing Tone Mapped Image Quality Index

“…With multiple TMOs at hand, a natural question is: which TMO produces the best quality tone mapped LDR image? This question could possibly be answered by subjective evaluation [10]- [13], which is expensive, time consuming, and perhaps most importantly, can hardly be used to guide automatic optimization procedures [14].…”