Closed cabouman closed 2 years ago
After changing detector pixel pitch from 0.25 to 1 (and increased dist_source_detector by 4x accordingly), I got nan values when performing qGGMRF recon in demo_mace3D.py:
dist_source_detector
demo_mace3D.py
************************** Iteration 0 (max. 20) ************************** * Cost = inf * Rel. Update = 0.0000000000e+00 % (threshold = 1.9999999553e-02 %) * RWFE = ||e||_W/||y||_W = -nan % (threshold = 0.0000000000e+00 %) * RUFE = ||e|| / ||y|| = 1.0000000000e+02 % (threshold = 0.0000000000e+00 %) * ---------------------------------------------------------------------------- * 1/M ||e||^2_W = inf = 1/0.0000000000 * weightScaler_value = 3.5543020203e-05 = 1/28134.9199218750 * ---------------------------------------------------------------------------- * voxelsPerSecond = 0.0000000000e+00 * time icd update = -1.3449177146e-01 s * ratioUpdated = 0.0000000000e+00 % * totalEquits = 0.0000000000e+00 ****************************************************************************** ************************** Iteration 1 (max. 20) ************************** * Cost = inf * Rel. Update = 1.0000000000e+02 % (threshold = 1.9999999553e-02 %) * RWFE = ||e||_W/||y||_W = -nan % (threshold = 0.0000000000e+00 %) * RUFE = ||e|| / ||y|| = 7.1919616699e+01 % (threshold = 0.0000000000e+00 %) * ---------------------------------------------------------------------------- * 1/M ||e||^2_W = inf = 1/0.0000000000 * weightScaler_value = 3.5543020203e-05 = 1/28134.9199218750 * ---------------------------------------------------------------------------- * voxelsPerSecond = 9.7759289062e+04 * time icd update = 7.2860288620e+00 s * ratioUpdated = 1.0000000000e+02 % * totalEquits = 1.0000000000e+00 ****************************************************************************** ************************** Iteration 2 (max. 20) ************************** * Cost = inf * Rel. Update = 6.1694126129e+01 % (threshold = 1.9999999553e-02 %) * RWFE = ||e||_W/||y||_W = -nan % (threshold = 0.0000000000e+00 %) * RUFE = ||e|| / ||y|| = 6.6280868530e+01 % (threshold = 0.0000000000e+00 %) * ---------------------------------------------------------------------------- * 1/M ||e||^2_W = inf = 1/0.0000000000 * weightScaler_value = 3.5543020203e-05 = 1/28134.9199218750 * ---------------------------------------------------------------------------- * voxelsPerSecond = 9.7759289062e+04 * time icd update = 6.7889585495e+00 s * ratioUpdated = 1.0000000000e+02 % * totalEquits = 2.0000000000e+00 ****************************************************************************** ************************** Iteration 3 (max. 20) ************************** * Cost = inf * Rel. Update = 3.0164772034e+01 % (threshold = 1.9999999553e-02 %) * RWFE = ||e||_W/||y||_W = -nan % (threshold = 0.0000000000e+00 %) * RUFE = ||e|| / ||y|| = 6.5200088501e+01 % (threshold = 0.0000000000e+00 %) * ---------------------------------------------------------------------------- * 1/M ||e||^2_W = inf = 1/0.0000000000 * weightScaler_value = 3.5543020203e-05 = 1/28134.9199218750 * ---------------------------------------------------------------------------- * voxelsPerSecond = 1.1405250000e+05 * time icd update = 6.3575387001e+00 s * ratioUpdated = 1.0000000000e+02 % * totalEquits = 3.0000000000e+00 ******************************************************************************
This only occurs when running demo_mace3D.py. demo_mace3D_fast.py works correctly, and reconstruction result makes sense.
demo_mace3D_fast.py
Updated demos
cabouman
commented
2 years ago
After changing detector pixel pitch from 0.25 to 1 (and increased
dist_source_detectorby 4x accordingly), I got nan values when performing qGGMRF recon indemo_mace3D.py:This only occurs when running
demo_mace3D.py.demo_mace3D_fast.pyworks correctly, and reconstruction result makes sense.