Benjamin_Loison/Robust_image_source_identification_on_modern_smartphones
1
0
Fork 0
Code Issues 68 Pull Requests Packages Projects Releases Wiki Activity

Use a low-pass filter denoiser #69

New Issue
Open
opened 2024-05-10 01:07:24 +02:00 by Benjamin_Loison · 6 comments
Benjamin_Loison commented 2024-05-10 01:07:24 +02:00
Owner
Copy Link

Someone asked this as the Stack Overflow question 54641616.

How to quantify the cutoff frequency (I mean how to deduce the circle diameter)?

Related to #70.

Someone asked this as [the Stack Overflow question 54641616](https://stackoverflow.com/q/54641616). How to quantify the cutoff frequency (I mean how to deduce the circle diameter)? Related to #70.
Benjamin_Loison added the
enhancement
medium priority
quick
 labels 2024-05-10 01:07:24 +02:00
Benjamin_Loison commented 2024-05-13 23:59:20 +02:00
Author
Owner
Copy Link

With Rafael 23/04/24:

mean_rafael_230424_low_pass_multiple_colors

With Rafael 23/04/24: ![mean_rafael_230424_low_pass_multiple_colors](/attachments/1869707b-f704-4054-bed9-9c3478c7c223)
mean_rafael_230424_low_pass_multiple_colors.png
39 MiB
Benjamin_Loison commented 2024-05-15 16:27:34 +02:00
Author
Owner
Copy Link

image

fft-then-ifft

fft-then-ifft_old

![image](/attachments/87a52a87-6272-459f-b318-75d2863c1ba2) ![fft-then-ifft](/attachments/637c255f-a81b-4d39-b784-43adbc38309b) ![fft-then-ifft_old](/attachments/8288eb0b-cf03-45a4-89ca-5794e8c9f664)
image.jpg
282 KiB
fft-then-ifft_old.png
148 KiB
fft-then-ifft.png
242 KiB
Benjamin_Loison commented 2024-05-15 17:17:07 +02:00
Author
Owner
Copy Link

The Stack Overflow answer 43346070 seems to be nice for square images. We can shift this square to the middle of the rectangle but does not seem very clean. Other answers to the associated question do not seem to treat the rectangle case.

Finally used the Stack Overflow answer 62949902.

[The Stack Overflow answer 43346070](https://stackoverflow.com/a/43346070) seems to be nice for square images. We can shift this square to the middle of the rectangle but does not seem very clean. Other answers to the associated question do not seem to treat the rectangle case. Finally used [the Stack Overflow answer 62949902](https://stackoverflow.com/a/62949902).
Benjamin_Loison commented 2024-05-15 17:35:57 +02:00
Author
Owner
Copy Link

fft-then-ifft

Interesting blurring without artifacts.

![fft-then-ifft](/attachments/4f315c7e-2446-44b2-826b-9463ba09e997) Interesting blurring without artifacts.
fft-then-ifft.png
146 KiB
Benjamin_Loison commented 2024-05-15 18:30:24 +02:00
Author
Owner
Copy Link
diff --git a/datasets/raise/utils.py b/datasets/raise/utils.py
index 6c699a0..973b4d1 100644
--- a/datasets/raise/utils.py
+++ b/datasets/raise/utils.py
@@ -73,6 +73,9 @@ def denoise(imageNpArray, denoiser):
 
             # Unsure that doing such *normalization* is appropriate.
             rectangleLowPassNpArray /= rectangleLowPassNpArray.max()
+            import matplotlib.pyplot as plt
+            plt.imsave('rectangle_low_pass.png', rectangleLowPassNpArray)
+            exit(1)
 
             # multiply both the images
             filtered = np.multiply(fft1, rectangleLowPassNpArray)

rectangle_low_pass

New corrected PRNU:

mean_rafael_230424_low_pass_multiple_colors

```diff diff --git a/datasets/raise/utils.py b/datasets/raise/utils.py index 6c699a0..973b4d1 100644 --- a/datasets/raise/utils.py +++ b/datasets/raise/utils.py @@ -73,6 +73,9 @@ def denoise(imageNpArray, denoiser): # Unsure that doing such *normalization* is appropriate. rectangleLowPassNpArray /= rectangleLowPassNpArray.max() + import matplotlib.pyplot as plt + plt.imsave('rectangle_low_pass.png', rectangleLowPassNpArray) + exit(1) # multiply both the images filtered = np.multiply(fft1, rectangleLowPassNpArray) ``` ![rectangle_low_pass](/attachments/132b8eb3-5be8-46f0-ac95-6631d7815699) New corrected PRNU: ![mean_rafael_230424_low_pass_multiple_colors](/attachments/51d41864-64b0-4488-9e15-8f500f0de883)
mean_rafael_230424_low_pass_multiple_colors.png
39 MiB
rectangle_low_pass.png
68 KiB
Benjamin_Loison commented 2024-05-30 17:49:55 +02:00
Author
Owner
Copy Link

Related to PRNU_extraction/issues/6. In this context I noticed this 2 very suspicious lines:

datasets/raise/utils.py Lines 99 to 101 in 7f7e78cafb
ifft2 = np.maximum(0, np.minimum(ifft2, 255))
imageDenoisedNpArray = ifft2.astype(np.uint8)

See PRNU_extraction/commit/1c9fa298defe3dd65b868606d06e1d661915d653.

Related to [PRNU_extraction/issues/6](https://codeberg.org/Benjamin_Loison/PRNU_extraction/issues/6). In this context I noticed this 2 very suspicious lines: https://gitea.lemnoslife.com/Benjamin_Loison/Robust_image_source_identification_on_modern_smartphones/src/commit/7f7e78cafb8f6d5807755b9801b0dcd2d8290776/datasets/raise/utils.py#L99-L101 See [PRNU_extraction/commit/1c9fa298defe3dd65b868606d06e1d661915d653](https://codeberg.org/Benjamin_Loison/PRNU_extraction/commit/1c9fa298defe3dd65b868606d06e1d661915d653).
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
bug
Something is not working
Context-Adaptive Interpolator
duplicate
This issue or pull request already exists
enhancement
New feature
epic
A task that is going to take more than a day to complete.
help wanted
Need some help
high priority
invalid
Something is wrong
left for future work
low priority
medium
A task that should take less than a day to complete.
medium priority
meta
question
More information is needed
quick
A task that should take less than two hours to complete.
wontfix
This won't be fixed
No Label
enhancement
medium priority
quick
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Benjamin_Loison/Robust_image_source_identification_on_modern_smartphones#69
No description provided.
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?