annotate.py

annotate.py contains two functions that you call from the python console.

Set up for a single image:

annotate_img(img_path, initials, size1, size2)

• Inputs

  • img_path - name and location of the image, example "B.jpg"
  • size1 - size of the 1st brush circle, in px
  • size2 - size of the 2nd brush circle, in px
  • initials - First name and last name initials of the person doing the labelling. Recorded next to every PMT feature ID line in the text file.

• Outputs

  • {image_name}.png - A map of all the features in binary - features labelled in blue are recorded as RGB (1,1,1), and features in red are recorded as RGB (2,2,2).
  • {image_name}.txt - A text file. Each line contains a PMT feature ID and its pixel coordinates on the image. The text file has the same name as the image.

---

Set up for directory of images with file structure for image segmentation. This function calls annotate_img() for every image in the specified directory, and then saves the output texts and masks in separate folders.

annotate_dir(img_dir, initials, size1, size)

• Inputs

  • img_dir - image directory. Images are contained inside this folder.
  • initials - First name and last name initials of the person doing the labelling. Recorded next to every PMT feature ID line in the text file.
  • size1 - size of the 1st brush circle, in px.
  • size2 - size of the 2nd brush circle, in px.

• Outputs (for every image in {img_dir})

  • {image_name}.png - A map of all the features in binary - features labelled in blue are recorded as RGB (1,1,1), and features in red are recorded as RGB (2,2,2).
  • {image_name}.txt - A text file. Each line contains a PMT feature ID and its pixel coordinates on the image. The text file has the same name as the image.

• Folder Structure

  • {img_dir}
    • Source images are stored here.
  • {img_dir}_texts
    • Output text files are stored here.
  • {img_dir}_masks
    • Output mask files are stored here.

---

User keypresses (make sure the image window is selected):

• click on image to grab pixel coordinates:
  • Left-click to label with red
  • middle-click to label with blue
• s to close image and exit program
• r to write coordinates (you will be prompted for the PMT feature ID if a PMT has not been selected yet)
• f to select new PMT
• n to increment the featureID number
• b to decrement the featureID number