(it is a Augmented reality project)

combined.py - this code will check each pixel and apply K Means clustering to find out a dominant colour . hex_color.py - will check for dominat colours and convert there RGB value to HEX value

i tried to delete the negate the pixel from a pertucular range but then the array falls short of 1 value and hence reshape() function from numpy does not work .

https://github.com/SPJasti/dominant_color_using_Kmeans/blob/master/notH_4.jpg sample image for testing

sample image's plot https://github.com/SPJasti/dominant_color_using_Kmeans/blob/master/plot1.png

sample image's - dominance in descending order https://github.com/SPJasti/dominant_color_using_Kmeans/blob/master/Dominant1.png

the original sample image i took had over 22500 pixel hence it took made a lot more iterations . i resized the image using open CV to 100x100 pixels now its 125% faster , sure there is datalot but in such case it can't be helped

i used this code for my project of handwash management system ( * soon whole project will be uploaded with a video on youtube *) even there i used resize because where i had to showcase the project internet connectivity was not available hence no *google firebase * just a raspberry pi doing all the computing . i have added flags so that if someone is willing to use the code you will understand what problem is that you are facing .

for notH_1.jpg

hist - [0.23961661 0.275449 0.1242441 0.16172985 0.19896044] will change for same image upon execution of code

cluster_centers - [[ 43.30203808 82.52354433 25.58764615] [169.38920051 145.72142132 118.39147208] [107.77231816 160.36423469 192.88628565] [ 40.01361515 97.08099931 133.81868084] [199.72480743 207.39182582 201.39465724]] will change for same image upon execution of code

Hex codes are: ['#a99176', '#2b5219', '#c7cfc9', '#276085', '#6a9fc0'] *won't change for same image upon execution of code any number of times *