Extract Color Data from Dictionary of Masks for Naive Bayes¶

This function is to extract color information from a dictionary of masks in order to get data for naive bayes functions. Collect pixel training data in Napari, rather than ImageJ for example.

plantcv.annotate.napari_naive_bayes_colors(img, maskdict, filename)

returns data frame

• Parameters:

  • img - RGB image to extract color information from
  • maskdict - dictionary of masks, output of napari_points_mask for example
  • filename - filename to save data, formatted to work with Naive Bayes segmentation

• Context:

  • This function is used to extract color information from an RGB image and a mask and convert data to be compatible with Naive Bayes training functions.

• Example use:

  • used in Napari Naive Bayes

import plantcv.plantcv as pcv 
import plantcv.annotate as pcvan
from plantcv import learn
import napari

# Create an instance of the Points class
img, path, name = pcv.readimage("./wheat.png")

# Should open interactive napari viewer
viewer = pcvan.napari_label_classes(img=img, classes=['background','healthy', 'rust', 'chlorosis'], size=4)

maskdict = pcvan.napari_points_mask(img, viewer)

nbdata = pcvan.napari_naive_bayes_colors(img=img, maskdict=maskdict, filename="./nbdata.txt")

learn.naive_bayes_multiclass(samples_file="./nbdata.txt", outfile="naive_bayes_test_pdfs.txt")

masks = pcv.naive_bayes_classifier(rgb_img=img, 
                                   pdf_file="./naive_bayes_test_pdfs.txt")

Output Data

Source Code: Here