Previous Contents Next
Previous : Reconstruction algorithm  Next : Reconstruction results 
 
Estimation results
 
 
 
Set of hyperparameters which cancels
the log-likelihood derivatives
(local minima)
see pictures below
 
 
 
Original image 
64 x 64		   $\lambda $=0.13, $\delta $=2 
SNR=27.5 dB		 $\lambda $=0.25, $\delta $=5 
SNR=28.9 dB		 $\lambda $=0.35, $\delta $=10 
SNR=28.4 dB
 
   
Observed image 
SNR=15.9 dB		   $\lambda $=0.66, $\delta $=30 
SNR=26.3 dB		 $\lambda $=0.91, $\delta $=50 
SNR=25.0 dB		  
   
       
 
Original image 
"plant" 128 x 128		   $\lambda $=0.19, $\delta $=2 
SNR=22.70 dB		 $\lambda $=0.33, $\delta $=5 
SNR=22.72 dB		 $\lambda $=0.49, $\delta $=10 
SNR=22.82dB
 
   
Observed image 
SNR=16.7dB		   $\lambda $=1.2, $\delta $=30 
SNR=22.79 dB		 $\lambda $=1.8, $\delta $=50 
SNR=22.73 dB		  
   
 Comparison between different reconstruction results
corresponding to a set of optimal hyperparameters
 
 
 
 
Signal to Noise Ratio (SNR) - "plant" image
around the optimal value of =0.19 (=2)
The best SNR really corresponds to the estimated.
 
Previous Next
André Jalobeanu - 24 Aug 1998