e20a0a5403
Rename net/hanshq/hello to com/robust_image_source_identification_on_modern_smartphones/camera
2024-05-02 01:22:43 +02:00
aca86e3344
Add and use $SOURCE_FOLDER and $APP_NAME
2024-05-01 23:03:09 +02:00
9af3b32820
Add datasets/android/compile.sh
2024-05-01 22:57:12 +02:00
92aa5a3558
Add datasets/android/
...
Based on https://www.hanshq.net/command-line-android.html .
See [Benjamin-Loison/android/issues/30](https://github.com/Benjamin-Loison/android/issues/30 ).
2024-05-01 22:44:36 +02:00
Benjamin Loison
53c3935bab
Clean some debugging
2024-04-30 06:48:17 +02:00
Benjamin Loison
7a807b91d8
#63 : Add debugging
2024-04-30 06:45:28 +02:00
Benjamin Loison
a8725e5e88
Move denoising testing images to work even if provide minColor and maxColor ( #63 )
2024-04-30 05:36:28 +02:00
Benjamin Loison
3c403dbed3
WIP ( #63 )
2024-04-30 05:31:14 +02:00
Benjamin Loison
b6425b426e
#63 : WIP
2024-04-30 05:01:05 +02:00
Benjamin Loison
319ca8fb60
Add attribute_source_camera.py
...
It is a modified copy of `split_and_compare_prnus_of_subgroups.py`.
2024-04-30 04:48:02 +02:00
Benjamin Loison
aa9f73e8a0
Fix random seed to improve reproducibility
2024-04-30 03:49:45 +02:00
Benjamin Loison
22fc4ce46b
Remove rescale for plt.imsave as do not seem necessary anymore
2024-04-30 03:36:37 +02:00
Benjamin Loison
20fcb247c6
Simplify loops
...
The idea was initially in the context of #62 to optimize execution
speed.
2024-04-30 03:36:04 +02:00
Benjamin Loison
ad5aa22ae6
Add missing import for saveNpArray and add a missing f-string
2024-04-30 03:35:30 +02:00
Benjamin Loison
c2a2dea2c7
Add a missing argument to getColorChannel in extract_noise.py
2024-04-30 03:22:03 +02:00
Benjamin Loison
31b979edb3
Override print to also print current time
2024-04-30 03:21:32 +02:00
Benjamin Loison
c9f75286bf
Make merge_single_color_channel_images_according_to_bayer_filter.py
...
virtual environment friendly
2024-04-30 03:20:44 +02:00
Benjamin Loison
ab9466c738
#62 : Optimize significantly mergeSingleColorChannelImagesAccordingToBayerFilter
2024-04-30 03:17:19 +02:00
Benjamin Loison
2d67c2bca1
Add and use rescaleRawImageForDenoiser and updateExtremes ( #59 )
2024-04-30 02:04:17 +02:00
Benjamin Loison
671a692114
#59 : Save rmss for futher investigations
...
For instance to plot a fair comparison, see [issues/59#issuecomment-1732](#59 (comment) ).
2024-04-30 01:04:33 +02:00
5647ab4019
#59 : Should make split_and_compare_prnus_of_subgroups.py compatible with RAW images
2024-04-27 23:13:48 +02:00
9b2f2281e4
Make merge_single_color_channel_images_according_to_bayer_filter.py more general
2024-04-27 22:57:29 +02:00
5ea0a862c0
Add and use getColorChannel
2024-04-27 21:49:32 +02:00
561ec9d1e0
Add and use escapeFilePath
2024-04-27 21:48:07 +02:00
b50f013234
Add and use isARawImage
2024-04-27 21:46:40 +02:00
a0fc38d023
Add and use getRawColorChannel to utils.py
2024-04-27 20:47:33 +02:00
Benjamin Loison
80a33428d6
Make split_and_compare_prnus_of_subgroups.py execution compatible with chaining shell commands
...
Otherwise it is stuck on `plt.show` until it is closed but we want to
chain with a notification command, so the notification is not
interesting in this case.
2024-04-27 19:36:45 +02:00
Benjamin Loison
9d8e5ddbec
Make split_and_compare_prnus_of_subgroups.py only use 2 images of RAM
...
quantity
2024-04-27 19:24:12 +02:00
Benjamin Loison
4ac75448a6
Correct in theory the range of color values
2024-04-27 17:54:16 +02:00
Benjamin Loison
643fee9f4a
Make split_and_compare_prnus_of_subgroups.py executable from terminal
...
Used to be only executable from Pyzo.
2024-04-27 16:55:03 +02:00
c8fe71444f
#59 : Leverage iterativeMean to not suffer of RAM quantity issue
2024-04-27 16:49:06 +02:00
Benjamin Loison
6f6721da8d
Make merge_single_color_channel_images_according_to_bayer_filter.py executable from terminal
...
Used to only run in Pyzo.
2024-04-27 16:20:02 +02:00
b0fa01c6c4
Remove potential chronological bias
2024-04-27 16:10:20 +02:00
e9dc28dc2a
Simplify and clean split_and_compare_prnus_of_subgroups.py
2024-04-27 16:06:47 +02:00
Benjamin Loison
ee7353cfde
Correct iterativeMean implementation for images
...
Otherwise get:
```
Traceback (most recent call last):
File "/home/benjamin/robust_image_source_identification_on_modern_smartphones/datasets/raise/./extract_noise.py", line 148, in <module>
treatImage(imageFileName, color = color)
File "/home/benjamin/robust_image_source_identification_on_modern_smartphones/datasets/raise/./extract_noise.py", line 114, in treatImage
estimatedPrnuIterativeMean.add(imageNoiseNpArray)
File "/home/benjamin/robust_image_source_identification_on_modern_smartphones/datasets/raise/utils.py", line 34, in add
self.mean = ((self.mean * self.numberOfElementsInMean) + element) / (self.numberOfElementsInMean + 1)
~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
TypeError: unsupported operand type(s) for *: 'NoneType' and 'int'
```
2024-04-27 15:58:29 +02:00
bd0b9aca94
Correct support for JPG images
2024-04-27 15:55:25 +02:00
e9038bdb76
Hide inappropriate figure title
2024-04-27 15:21:57 +02:00
af947dbf65
Add articles/Euclidean distance - Wikipedia/
2024-04-27 15:08:56 +02:00
5baabae99d
Make mean initialization more general in iterativeMean
2024-04-26 15:56:10 +02:00
1ea1bde910
Leverage iterativeMean in extract_noise.py
2024-04-26 14:25:08 +02:00
674f87d480
Remove unnecessary iterativeMean parentheses and rename addElement to add
2024-04-26 14:22:59 +02:00
7af96d6bf6
Add iterativeMean to utils
...
Verified with:
```py
myIterativeMean = iterativeMean()
print(myIterativeMean.mean)
myIterativeMean.addElement(2)
print(myIterativeMean.mean)
myIterativeMean.addElement(3)
print(myIterativeMean.mean)
```
Note that cannot simplify as follows in the general case of using images as
`element`s for instance:
```diff
diff --git a/datasets/raise/utils.py b/datasets/raise/utils.py
index fba5523..a116edb 100644
--- a/datasets/raise/utils.py
+++ b/datasets/raise/utils.py
@@ -27,12 +27,9 @@ def denoise(imageNpArray, denoiserName):
return imageDenoisedNpArray
class iterativeMean():
- mean = None
+ mean = 0
numberOfElementsInMean = 0
def addElement(self, element):
- if self.mean is None:
- self.mean = element
- else:
- self.mean = ((self.mean * self.numberOfElementsInMean) + element) / (self.numberOfElementsInMean + 1)
+ self.mean = ((self.mean * self.numberOfElementsInMean) + element) / (self.numberOfElementsInMean + 1)
self.numberOfElementsInMean += 1
\ No newline at end of file
```
2024-04-26 14:16:35 +02:00
7684e1408d
Remove unused imports in split_and_compare_prnus_of_subgroups.py
2024-04-26 13:58:19 +02:00
9c469cf67a
Leverage utils.py denoise in split_and_compare_prnus_of_subgroups.py
2024-04-26 13:58:00 +02:00
9d2903af60
Move denoise from extract_noise.py to utils.py
...
See next commit leveraging it in `split_and_compare_prnus_of_subgroups.py`
2024-04-26 13:57:18 +02:00
6a93b04686
Remove already used previous temporary analysis commit code
2024-04-26 13:49:31 +02:00
b52fff9deb
Add commented code to enforce merging Numpy arrays with a greyscale
...
colormap.
2024-04-26 13:48:44 +02:00
3ffcd33efe
Add Gaussian filter to blur images when using mean as denoiser
...
See
https://docs.scipy.org/doc/scipy-1.13.0/reference/generated/scipy.ndimage.gaussian_filter.html#ba1e4df0-d4be-47c2-81e5-059286837a2b .
2024-04-26 04:39:06 +02:00
9641641b61
Remove commented code from analyze_bayer_filter_mean_denoiser.py
2024-04-26 04:13:08 +02:00
3e8b607b97
Add analyze_bayer_filter_mean_denoiser.py
2024-04-26 04:12:13 +02:00
