Commit Graph

246 Commits

Author SHA1 Message Date
85bc972c54
Clean completely analyze_fft_ellipses.py 2024-05-22 20:13:20 +02:00
80ac1f1bf3
Clean analyze_fft_ellipses.py 2024-05-22 19:53:20 +02:00
ffd71ce5b8
Add analyze_fft_ellipses.py 2024-05-22 19:19:09 +02:00
7a002f8b0d
Add 1 to np.log10 argument 2024-05-22 16:43:17 +02:00
99fb31bbd1
Use same color scale 2024-05-16 09:55:36 +02:00
da5b19f047
Clean image_to_fourier.py 2024-05-16 01:39:06 +02:00
eb33def18c
Add a missing : and imports 2024-05-16 00:33:29 +02:00
Benjamin Loison
f719ff767a
#66: Stop execution if Pearson correlation is negative 2024-05-15 20:15:20 +02:00
Benjamin Loison
60a9dfe89c
Update forgotten to commit scripts 2024-05-15 20:00:44 +02:00
Benjamin Loison
6e7e06b85c
Clean attribute_source_camera.py 2024-05-15 19:57:56 +02:00
Benjamin Loison
e3e3d44ca9
#62: Comment loading images to memory in favor of cropping 2024-05-15 19:57:04 +02:00
6a9900df91
#66: Add Distance Enum to choose between ROOT_MEAN_SQUARE and PEARSON_CORRELATION 2024-05-15 19:51:30 +02:00
Benjamin Loison
7ed6ff5adb
Add RESOLUTION support for Denoiser.MEAN in extract_noise.py 2024-05-15 19:20:19 +02:00
Benjamin Loison
ec4657eda8
Clean low pass denoiser 2024-05-15 18:21:11 +02:00
Benjamin Loison
95796b53fc
#69: Remove artifacts from low pass denoiser
See [Robust_image_source_identification_on_modern_smartphones/issues/69#issuecomment-1892](#69 (comment)).
2024-05-15 17:41:19 +02:00
Benjamin Loison
462bd92509
Make production ready benchmark_load_part_of_images.py 2024-05-15 16:23:39 +02:00
f7c08849dc
Update old merge_with_same_color_scale_single_color_channel_images_according_to_bayer_filter.py 2024-05-15 15:43:16 +02:00
8ec292ba11
Add old Termux try 2024-05-15 15:42:25 +02:00
dac9ce9d50
Update remove_period_patterns.py 2024-05-15 15:39:11 +02:00
7dcd1f3331
Update remove_period_patterns.py 2024-05-14 11:40:56 +02:00
337a6eca84
Add titles 2024-05-14 02:12:54 +02:00
038f5ffc00
Add remove_period_patterns.py 2024-05-14 02:03:46 +02:00
b7c494579c
Add old get_ft_prnu.py 2024-05-14 01:19:57 +02:00
810bde8e35
Add fft scripts 2024-05-14 01:19:05 +02:00
4bab354a95
Clean fft.py 2024-05-14 00:24:02 +02:00
451bd55219
Add fft.py 2024-05-14 00:17:29 +02:00
Benjamin Loison
ce64450101
Load lazily images (#62) 2024-05-13 20:06:47 +02:00
Benjamin Loison
10f8766793
First try lazy load 2024-05-13 19:50:54 +02:00
Benjamin Loison
f85e63b6e0
#62: First lazy load try
```
OPERATION = <Operation.LOAD_NPY: 3>
RESOLUTION = 100
Image:   0%|▍                                                                                            | 1/200 [00:00<00:00, 5084.00it/s]
Camera:   0%|                                                                                                        | 0/2 [00:00<?, ?it/s]
Traceback (most recent call last):
  File "/home/benjamin/robust_image_source_identification_on_modern_smartphones/datasets/raise/./benchmark_load_part_of_images.py", line 43, in <module>
    rawColorChannels = np.load(numpyFilePath, mmap_mode = 'r', allow_pickle = True)
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/benjamin/venv/lib/python3.11/site-packages/numpy/lib/npyio.py", line 453, in load
    return format.open_memmap(file, mode=mmap_mode,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/benjamin/venv/lib/python3.11/site-packages/numpy/lib/format.py", line 932, in open_memmap
    raise ValueError(msg)
ValueError: Array can't be memory-mapped: Python objects in dtype.
```
2024-05-13 18:30:28 +02:00
Benjamin Loison
ac32abede1
Save all colors 2024-05-13 18:22:25 +02:00
Benjamin Loison
36b17eb3cc
#62: Add benchmark_load_part_of_images.py 2024-05-13 17:52:49 +02:00
Benjamin Loison
fb0f78e069
Add the ability to hardcode resolution 2024-05-13 16:27:53 +02:00
Benjamin Loison
86093a40c0
Load training images to memory to ease implementation and make execution
faster
2024-05-13 16:26:58 +02:00
Benjamin Loison
54178c1101
#72: Verify actual data type pointer 2024-05-13 16:10:11 +02:00
Benjamin Loison
50058d5d2e
#72: Verify pointer before actually implementing the correct second choice 2024-05-13 15:55:59 +02:00
Benjamin Loison
dbfc2756b2
Remove commented code 2024-05-13 15:21:30 +02:00
Benjamin Loison
31704c6e78
#72: Implement, possibly incorrectly, first choice 2024-05-13 15:20:17 +02:00
Benjamin Loison
5e6702e007
#63: use scipy.stats.pearsonr 2024-05-13 12:51:26 +02:00
Benjamin Loison
1a910d2e2c
Add a try with numpy.corrcoef 2024-05-13 12:39:59 +02:00
Benjamin Loison
6ed57bc477
Precise purpose of a part of code 2024-05-13 03:13:58 +02:00
Benjamin Loison
3f7b6bf21a
Replace _ with for end-user color channel name 2024-05-13 02:56:27 +02:00
Benjamin Loison
b2ffc20473
Add comments to attribute_source_camera.py 2024-05-13 02:48:22 +02:00
Benjamin Loison
226936020e
Remove comments of {min,max}Color 2024-05-13 02:02:51 +02:00
934add5d91
Add boxplot for the specified denoisers 2024-05-10 00:14:30 +02:00
73fb61b16d
Correct missing denoiser usage 2024-05-10 00:13:47 +02:00
8523c02517
Fix #68: Use same color map for different images to more accurately compare them 2024-05-10 00:03:08 +02:00
3e332c19d8
Remove unnecessary from PIL import Image 2024-05-10 00:02:44 +02:00
Benjamin Loison
f1495f2ba0
Add and use PREDICT_ONLY_ON_WHOLE_TRAINING_SET 2024-05-03 13:55:32 +02:00
Benjamin Loison
612f4119fd
Determine the mean of camera images only on the training set 2024-05-03 03:54:26 +02:00
Benjamin Loison
d7f7728211
Add Denoiser Enum 2024-05-03 03:49:15 +02:00