20260113_202254_silicon and 20260113_203132_silicon: Tried 10 percent sampling on 'silicon' image, not as good as CIGS image.20260113_203132_silicon: Increased eta to 1e4 compared to 1e2 in 20260113_202254_silicon, worse results, so eta can stay at 1e2.20260113_204459_silicon: Tried 20 percent sampling, still not good. Probably the silicon image contains a lot more details than CIGS image, so higher sampling ratio is needed.20260113_205129_silicon_256x256: Tried 25 percent sampling, still artifacts20260113_214748_silicon_512x512: Tried 25 percent sampling on essentially the same silicon image but higher resolution (512x512) (taken from Andrew's paper). Results are much better. This suggests that higher resolution requires lower sampling ratio.20260113_222052_silicon_512x512: Tried 20 percent sampling, visualise with color map 'afmhot' with max at 0.8 to recreate Andrew's paper style.20260113_223817_silicon_512x512_color_and_zoom: Same as 20260113_222052_silicon_512x512 but added zoom-inset and color bar using PlotHelper class.
