Plot Delta E

This function creates an interactive chart visualizing per-chip Delta E (color difference) values from a color checker card. Each bar or point is colored by a standard interpretation category: green (<1) indicates imperceptible differences, progressing through yellow (<2, <10) to orange and red (<49, >49) for increasingly noticeable differences. Reference lines at the category boundaries are drawn for quick visual assessment. Standard chip color swatches from the target color matrix are displayed below the x-axis to aid chip identification.

When passed a single numpy.ndarray of Delta E values, a bar chart is produced for that image. When passed a directory path or list of image paths, the function reads each image, computes Delta E, and returns a layered boxplot with jittered points so per-image variation is visible alongside the distribution.

plantcv.qc.plot_deltaE(source, n=20, ext="png", **kwargs)

returns chart, an altair.LayerChart

  • Parameters:

    • source - one of:
      • numpy.ndarray of per-chip Delta E values shaped to match the color card layout (e.g., (6, 4) for a 24-chip Macbeth card or (3, 5) for a 15-chip AstroBotany card), as returned from pcv.transform.deltaE. A bar chart is produced.
      • str path to a directory of images (walked recursively). A boxplot is produced.
      • list of file path strings to individual images. A boxplot is produced.
    • n - (default: 20) maximum number of images to read when source is a directory path
    • ext - (default: "png") file extension filter when source is a directory path
    • **kwargs - additional arguments forwarded to pcv.transform.deltaE: color_chip_size, roi, adaptive_method, block_size, radius, min_size, aspect_ratio, solidity
  • Context:

    • Used to evaluate the quality of color calibration by visualizing how closely observed chip colors match their expected values. Lower Delta E values indicate better color fidelity. The single-image bar chart is best used during workflow development to inspect calibration results interactively. The multi-image boxplot is useful for assessing calibration consistency across an experiment or imaging session.
  • Example use:

    • Below

Dataset image:

Screenshot

from plantcv import plantcv as pcv

# Single image — bar chart
de_matrix = pcv.transform.deltaE(rgb_img=img, color_chip_size="classic")
chart = pcv.qc.plot_deltaE(source=de_matrix)
# Calculate Delta E values for each chip relative to the standard color matrix
de_matrix2 = pcv.transform.deltaE(rgb_img=img, color_chip_size="classic")
# Plot the Delta E values
chart = pcv.qc.plot_deltaE(source=de_matrix2)

Delta E bar chart:

Screenshot

# Directory of images — boxplot with jittered points
chart = pcv.qc.plot_deltaE(source="/path/to/images", n=50, ext="png")

# List of image paths — boxplot with jittered points
image_paths = ["/path/to/img1.png", "/path/to/img2.png"]
chart = pcv.qc.plot_deltaE(source=image_paths)

Delta E boxplot:

Screenshot

Source Code: Here