how to tune the paramteres, the reconstruction is not good

[argman]

import os
import numpy as np
import matplotlib.pyplot as plt
import svmbir

"""
Fan beam demo
"""

# Simulated sinogram parameters
geometry = 'fan-curved'
dist_source_detector = 1000.0
delta_channel = 1.0
magnification = 2.0
num_views = 512
num_channels = 256
angles = np.linspace(-np.pi, np.pi, num_views, endpoint=False)

# Reconstruction parameters
img_size = 256
snr_db = 30.0
sharpness = 0.0
T = 0.1
p = 1.2

# Generate phantom with a single slice
# phantom = svmbir.phantom.gen_shepp_logan(img_size,img_size)
phantom = np.load('sp.npy')
vmin, vmax = phantom.min(), phantom.max()
phantom = (phantom - vmin) / (vmax - vmin)

phantom = np.expand_dims(phantom, axis=0)
sino = svmbir.project(phantom, angles, num_channels, geometry=geometry,
    dist_source_detector=dist_source_detector, magnification=magnification,
    delta_channel=delta_channel)

# Perform MBIR reconstruction
recon = svmbir.recon(sino, angles, num_rows=img_size,
    num_cols=img_size, T=T, p=p, sharpness=sharpness, snr_db=snr_db,
    geometry=geometry, dist_source_detector=dist_source_detector,
    magnification=magnification, delta_channel=delta_channel)

# Compute Normalized Root Mean Squared Error
print(np.mean(np.abs(recon[0] - phantom[0])), np.max(phantom[0]))
nrmse = svmbir.phantom.nrmse(recon[0], phantom[0])

#plt.ion()
plt.figure(); plt.imshow(phantom[0] * (vmax - vmin) + vmin,vmin=-80,vmax=160); plt.colorbar()
plt.title('Shepp Logan Phantom')
plt.savefig('output/shepp_logan_phantom.png')

plt.figure(); plt.imshow(np.squeeze(sino).T); plt.colorbar()
plt.title('Sinogram')
plt.savefig('output/shepp_logan_sinogram.png')

plt.figure(); plt.imshow(recon[0] * (vmax - vmin) + vmin,vmin=-80,vmax=160); plt.colorbar()
plt.title(f'Reconstruction, nmrse={nrmse:.3f}')
plt.savefig('output/shepp_logan_recon_fanbeam.png')

print("Close figures to continue")
plt.show()

sp.zip