-
Notifications
You must be signed in to change notification settings - Fork 5
Generation Outputs
All outputs from the program are stored in a directory called output. In it, you will find several files which I will go over in detail.
sin.csv and sin_smooth.csv are two column csv files. The first column contains the frame number. The second column contains the average pixel value across that frame. The sin_smooth.csv file contains the data from sin.csv with wave smoothing applied to it. The sin csv files were included so that the user can check if the drop generation rate seems reasonable. The drop generation rate is the period of the sin wave graphed from the data. If you do not trust the program output, check the sin_smooth.csv file manually for holes or double peaks. Cross check this with the frame information from raw to see if this is reasonable.
The file drop_data.csv contains the information that can be found from the standard output, but in file form. This is to make it easier for any developers who want to create an automatic pipeline using this script. The file contains (in order):
- Average area of the droplets in pixels
- Average area of the droplets in micrometers
- Standard deviation of the droplets in pixels
- Standard deviation of the droplets in micrometers
- Drop Generation Rate
drop_data_raw.csv contains the area of each droplet in a comma separated list. You can use this file to check if the answers seem reasonable, or if the standard deviation/mean is thrown off by a single value.
For most checking though, you are better off viewing the images in the cleaned directory to see if the drawn areas seem reasonable.
For each frame, the image is cropped to the bounding box and sent to output/raw.
These are the modified images that the pixel average value is calculated from.
If there seems to be an error with your areas or if the sin wave is too flat or too noisy, check these images. If they seem off, there is probably a problem with the edge detection parameters. If the tolerance is too low, these images will have a ton of noise. If it is too high, the droplets will be halfway formed or perhaps even blank.
This folder is provided as a kind of extension to the edge folder. The white area is the convex hull of the contours of the raw image. The blue bounding box is used to find the droplet diameter. If there are problems with the area, it should be very clear from this folder.
The analysis software parses this output in a clear way and allows you fix problems on the fly. The outputs on this folder are left if you want to generate your own graphs and figures for publication or if you want to statistically analyze the data in a more advanced way.