Rgb_to_colour_name = closest_colour(requested_colour) Rgb_to_colour_name = actual_name = webcolors.rgb_to_name(requested_colour) This thread for the closest colour algorithm here def get_colour_name(requested_colour): Here, while looping through each colour extracted, we call a new function, using the webcolors module. The first step to convert RGB to colour names is to edit our colour extraction function. The main improvement which we can make is assign 'names' to our RGB values which will allow us to combined multiple shades of the same colour into one. Larger image files will take longer to process - a compression function is required to optimise this step.Slightly different shades of the same colour will be counted separately which can skew certain images.The same RGB colour could be displayed multiple times - here, the proportions should be summed.The output here works well for the most part, however there are improvements which can be made: Plt.savefig('exports/' + folder + '.jpg') Img = /255, colour_extracts/255, colour_extracts/255)]]įinally, we plot the img onto our figure, number the colour on our palette and save the plotted image (our colour palette) to a. Print(f"Number of images processing: """) TEMP_FOLDER_IMAGES = os.listdir(MAIN_FOLDER + folder)įOLDER_IMAGES = Proportions = def extract_colour_values(folder): These are all saved into a list of lists: import os The following function takes a folder from your directory, loads each image file and extracts the RGB and proportion values. Proportion (%) of the image which each RGB value takes up.
The first step is to extract the most dominant colours from each image.
Python color palette from image install#
pip install matplotlib - we are using matplotlib to view and save our colour palettes.pip install webcolors - this module provides HTML and CSS colour definitions, allowing us to convert RBG and otherĬolour formats their closest colour names.pip install colorgram.py - colourgram is a simple module which allows us to get the RGB and HSL values from an image.There are a few Python 3 packges requried for this tool:
0 kommentar(er)
