A script that determines if 3D images are noisy or not based on the number of non-zero pixels in the image. Reads in the images from their locations given in a text file, then outputs if each image is clean/not noisy (True) or noisy (False) in an output text file.
To run the script, the expected format of the command is
python noise.py [input text file] [threshold*] [2D or 3D] [output text file]
The input text file holds the paths to all the images to check for noise.
The threshold is the cutoff between a clean/not noisy image and a noisy one. We've have used 1.2 x 10^8, but 1.4 or 1.6 would work as well.
2D or 3D tells the script what type of images it should be expecting.
The output text file will hold the output for each image in the input text file.
- NOTE: The threshold input is implicitly multiplied by 10^8 (or 100 million)
In the input text file, the script expects the format to have one image file path per line. For example,
folder1/image_file_1.tiff
folder2/image_file_2.tiff
The output file will be written as
[image file name],True or False
True indicates a clean image and False indicates a noisy image
- Add in the ability to choose between using the non-zero count method or the sum of pixel intensity method to determine the cutoffs
- If wanted, add print statements or a debug mode to show what the script is doing
- See if there are ways to make the script run faster
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent