A beam hardening and dispersion correction for x‐ray dark‐field radiography. Issue 6 (10th May 2016)

• Record Type: Journal Article
• Authors: Pelzer, Georg; Anton, Gisela; Horn, Florian; Rieger, Jens; Ritter, André; Wandner, Johannes; Weber, Thomas; Michel, Thilo
• Is Part Of: Medical physics. Volume 43:Issue 6(2016)Part 1
• Page Start: 2774
• Page End: 2779
• Publication Date: 2016-05-10
• Subjects: biological tissues -- diagnostic radiography -- mammography -- medical image processing -- phantoms; Mammography; Biological material, e.g. blood, urine; Haemocytometers -- Digital computing or data processing equipment or methods, specially adapted for specific applications -- Image data processing or generation, in general; x‐ray imaging -- phase‐contrast -- dark‐field -- beam hardening; Visibility -- Diffraction gratings -- Medical X‐ray imaging -- Mammography -- Dispersion -- Tissues -- X‐ray detectors -- Wave attenuation -- X‐ray imaging; Medical physics -- Periodicals; Medical physics; Geneeskunde; Natuurkunde; Toepassingen; Biophysics; Periodicals; Periodicals; Electronic journals; 610.153
• DOI: 10.1118/1.4948671
• Languages: English
• ISSNs: 0094-2405

Abstract

Abstract : Purpose: X‐ray dark‐field imaging promises information on the small angle scattering properties even of large samples. However, the dark‐field image is correlated with the object's attenuation and phase‐shift if a polychromatic x‐ray spectrum is used. A method to remove part of these correlations is proposed. Methods: The experimental setup for image acquisition was modeled in a wave‐field simulation to quantify the dark‐field signals originating solely from a material's attenuation and phase‐shift. A calibration matrix was simulated for ICRU46 breast tissue. Using the simulated data, a dark‐field image of a human mastectomy sample was corrected for the finger print of attenuation‐ and phase‐image. Results: Comparing the simulated, attenuation‐based dark‐field values to a phantom measurement, a good agreement was found. Applying the proposed method to mammographic dark‐field data, a reduction of the dark‐field background and anatomical noise was achieved. The contrast between microcalcifications and their surrounding background was increased. Conclusions: The authors show that the influence of and dispersion can be quantified by simulation and, thus, measured image data can be corrected. The simulation allows to determine the corresponding dark‐field artifacts for a wide range of setup parameters, like tube‐voltage and filtration. The application of the proposed method to mammographic dark‐field data shows an increase in contrast compared to the original image, which might simplify a further image‐based diagnosis.