Quantitative Visually Lossless Compression Ratio Determination of JPEG2000 in Digitized Mammograms

Abstract

The current study presents a quantitative approach towards visually lossless compression ratio (CR) threshold determination of JPEG2000 in digitized mammograms. This is achieved by identifying quantitative image quality metrics that reflect radiologists' visual perception in distinguishing between original and wavelet-compressed mammographic regions of interest containing microcalcification clusters (MCs) and normal parenchyma, originating from 68 images from the Digital Database for Screening Mammography. Specifically, image quality of waveletcompressed mammograms (CRs, 10:1, 25:1, 40:1, 70:1, 100:1) is evaluated quantitatively by means of eight image quality metrics of different computational principles and qualitatively by three radiologists employing a five-point rating scale. The accuracy of the objective metrics is investigated in terms of (1) their correlation (r) with qualitative assessment and (2) ROC analysis (A z index), employing pooled radiologists' rating scores as ground truth. The quantitative metrics mean square error, mean absolute error, peak signal-to-noise ratio, and structural similarity demonstrated strong correlation with pooled radiologists' ratings (r, 0.825, 0.823, −0.825, and −0.826, respectively) and the highest area under ROC curve (A z , 0.922, 0.920, 0.922, and 0.922, respectively). For each quantitative metric, the highest accuracy values of corresponding ROC curves were used to define metric cut-off values. The metrics cut-off values were subsequently used to suggest a visually lossless CR threshold, estimated to be between 25:1 and 40:1 for the dataset analyzed. Results indicate the potential of the quantitative metrics approach in predicting visually lossless CRs in case of MCs in mammography.

Evaluating image quality measures to assess the impact of lossy data compression applied to climate simulation data

“…Given these results, it is interesting to note that the SSIM threshold used in [BXH∗17] was chosen as 0.98 based on suggestions from the medical literature (e.g, [GKS∗12a], [Weg10]). Using a SSIM threshold of 0.98, the estimated proportion different for this threshold value is 0.9588 (0.9740, 0.9354), where the values in parenthesis provide the 95% confidence bounds for the estimate.…”

“…The authors in [BXH∗17] indicate that the SSIM threshold chosen was largely inspired by suggestions from the medical imaging field, where lossy compression is typically used to reduce unmanageable data volumes (e.g., [CP16, KR12, Wan11, KS06]). Indeed, the loss of critical information (a concern shared by climate scientists) is understandably a concern with procedures such as computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound imaging (e.g., [KBMG13, KMLH10, GKS∗12b]). However, an acceptable SSIM threshold varies by application, and the authors in [BXH∗ 17] acknowledge that further research is required to confidently select one for CESM images.…”

DSSIM: a structural similarity index for floating-point data

“…The constant values are of particular interest when applying directly to floating-point simulation data, as discussed in the next section. Another implementation detail is the local window (or patch) for which the per-pixel SSIM value statistics (mean, variance and covariance) are computed in (8). The recommendation in [1] is an 11 × 11 window with a Gaussian filter kernel.…”

“…For example, in the medical imaging field, images are typically compressed to reduce unmanageable data volumes, but clearly the potential loss of critical information (i.e., details needed to make an accurate diagnosis) is a concern. The SSIM has been advocated as a means of evaluating compressed-medicalimage quality in many studies (e.g., see [8]- [12]).…”

Limitations of the SSIM quality metric in the context of diagnostic imaging

“…These databases are important tools used to evaluate the performance of image quality assessment (IQA) metrics in the context of streamed videos or pictures displayed on web pages, but their results may not translate well to text images, graphics or diagnostic images. Some studies involving diagnostic images and trained radiologist have shown SSIM to be either on par [6,7] with or better [8] than MSE, but they have not explored its limitations.…”