2 Commits

  1. 51
      README.md
  2. 1
      colonycounter.py
  3. 6
      fileutils.py
  4. BIN
      readme_img/Filter_Colony.gif
  5. BIN
      readme_img/Remove_Colony.gif
  6. 2
      requirements.txt

51
README.md

@ -0,0 +1,51 @@
[MIT License](./LICENSE)
# Colony Counter
A simple Utility to count and measure bright spots on dark images 😸
## Features
- Image transformation results are cached and incremental.
- Provides an ugly user interface with intuitive key binds.
- Batch process/export image analysis results.
- Adjust colony recognition and filtering through a couple of simple sliders.
![](./readme_img/Filter_Colony.gif)
- Remove false positives by clicking on them.
![](./readme_img/Remove_Colony.gif)
## Getting Started
- Install python (>=3.8 recommended... I made sure to explicitly use ordered dictionaries but just in case).
- Install dependencies in whichever way you prefer.
- Create the folders "in", "out" and "cache" (or run colonycounter.py).
- Adjust queue functions in source code to suit your needs (and read through [Caveats](#caveats)).
- Copy/Move your images into the "in" folder.
- Start the program by running the colonycounter.py file.
## Key Binds
| Key | Function |
|-----------------|----------------------------------------------------------------------------------------|
| q, left arrow | load previous image |
| w, right arrow | load next image |
| a | apply changes (always do this before switching images, otherwise changes will be lost) |
| r | reset changes that have not yet been applied |
| l | load saved settings from disk |
| s | save currently applied settings to disk |
| left mouse btn | add colony to list of ignored colonies |
| right mouse btn | remove colonies in the vicinity from ignore list |
| c | clear list of ignored colonies |
| e | batch process and export all images |
## Caveats
- The current processing queue expects 16bit gray scale images with the file extension .tif (tagged image file).
- watershed function does not work on 16 bit images... so there are some hard coded values to bring measurements back to 16bit brightness values
- this means a loss of resolution
- To get hour and culture type values the following folder structure is expected:
- "$DIRIN/OPTIONAL\_FOLDERS/$HOUR/$CULTURETYPE/image.tif"
- $HOUR = integer followed by the letter h
- $CULTURETYPE = whatever string identifies your culture type
- Code documentation is currently lacking
- Not all types are properly annotated

1
colonycounter.py

@ -122,7 +122,6 @@ def main():
[f.opqueue.__init__([normalize, to8bit, rollingball, wat]) for f in files]
[f.apply(dslice(f.opqueue,None,-1)) for f in files]
# [(f.apply(f.opqueue), f.save(True, f.opqueue)) for f in files]
window.init(files)
window.reset()
return True

6
fileutils.py

@ -53,7 +53,7 @@ class FILE:
self.meta = pd.DataFrame(None if self.data[1]["data"][1:] == [] else self.data[1]["data"][1:], columns=self.data[1]["data"][0])
self.meta.insert(loc=0, column="Label", value="".join([self.fname, self.fext]))
spath = self.path.split(path.sep)
if len(spath) > 2 and spath[-1] in ["BFP", "YFP"] and spath[-2][:-1].isdigit():
if len(spath) > 2 and spath[-2][:-1].isdigit():
hour, culture = self.path.split(path.sep)[-2:]
hour = int(hour[:-1])
else:
@ -82,11 +82,7 @@ class FILE:
if self.outlines is not None:
imwrite(path.join(DIROUT, self.path, f"{self.fname}.outlines.jpg"), self.outlines)
if self.meta is not None:
print(path.join(DIROUT, self.path, f"{self.fname}.csv"))
self.meta.to_csv(path.join(DIROUT, self.path, f"{self.fname}.csv"), index=False)
# [l.insert(1, self.getName(False, {})) for l in self.meta["data"][1:]]
# with open(path.join(DIROUT, self.path, f"{self.fname}.csv"), "w") as f:
# [f.write(",".join([str(e) for e in l]) + "\n") for l in self.meta["data"]]
else:
if not path.exists(path.join(DIRCACHE, self.path)):
makedirs(path.join(DIRCACHE, self.path))

BIN
readme_img/Filter_Colony.gif

After

Width: 1028  |  Height: 1023  |  Size: 174 KiB

BIN
readme_img/Remove_Colony.gif

After

Width: 512  |  Height: 512  |  Size: 130 KiB

2
requirements.txt

@ -1,5 +1,5 @@
numpy
opencv-python == 4.3.0
pythreshold
scikit-image
scikit-image >= 0.18.1
scipy
Loading…
Cancel
Save