view article
Figure 6
Heatmap illustrating the performance of center detection methods on the single-crystal dataset D2. Each row corresponds to an individual diffractogram [15 patterns per material: (1) Au, (2) TbF3, (3) Fe3O4, (4) LaF3 and (5) GdF3]. Each column represents a center detection method (manual = three-point detection, phase = phase cross-correlation, ccorr = autocorrelation, intensity = maximum intensity detection, curvefit = pseudo-Voigt profile fitting, hough = Hough transform). Hatched columns indicate a method's performance on pre-processed diffractograms (edge detection using a Sobel operator). The color bar on the right indicates the performance scale: the lightest color (yellow) indicates the most accurate detection, while the darkest color (black) signifies failure or substantial inaccuracy in center location. The groups (1)–(5) correspond to various diffractogram qualities, as explained in Section 2.3[link], Table 4 and Table S3. The quality is as follows: (1) perfect, (2) good, (3) bad, (4) poor and (5) intermediate.

Journal logoJOURNAL OF
APPLIED
CRYSTALLOGRAPHY
ISSN: 1600-5767
Volume 59| Part 3| June 2026| Pages 845-857
https://doi.org/10.1107/S1600576726002384
Follow J. Appl. Cryst.
Sign up for e-alerts
E-alerts
Follow J. Appl. Cryst. on Twitter
Twitter
Follow us on facebook
Facebook
Sign up for RSS feeds
RSS