Synthetic Results: Station Scene
PSNR of each frame in the input sequence and in the results of the 3 methods in compaison. Our method performs best in terms of preserving fine structural details.
Frame 20
Please move your mouse cursor over the small images to toggle between the results.
| Constant exposure (1/60) |
Noisy input PSNR = 25.08 |
CBM3D PSNR = 35.44 |
Liu and Freeman PSNR = 35.92 |
Our algorithm PSNR = 38.94 |
Ground truth |
|---|---|---|---|---|---|
|
|
|
|
|
|
Constant exposure time |
|||||
Our result is comparable to CBM3D for this frame. Lines on the wall and details in the reflection are preserved better in comparison to Liu and Freeman's method. Difference images with respect to the ground truth are shown below for the full frame.
| Ground truth | CBM3D | Liu and Freeman | Our algorithm |
|---|---|---|---|
|
|
|
|
Ground truth image |
|||
Frame 31
| Constant exposure (1/60) |
Noisy input PSNR = 25.58 |
CBM3D PSNR = 35.92 |
Liu and Freeman PSNR = 35.90 |
Our algorithm PSNR = 38.51 |
Ground truth |
|---|---|---|---|---|---|
|
|
|
|
|
|
Constant exposure time |
|||||
Our algorithm performs significantly better than the others in this frame. When we toggle between each result and the ground truth, structured objects are sharper and better preserved in our result. Difference images are shown below for the full frame.
| Ground truth | CBM3D | Liu and Freeman | Our algorithm |
|---|---|---|---|
|
|
|
|
Ground truth image |
|||
