Color Detection and segmenataion image with threshholds for non uniform colors

here you can see some outputs of the programe. In each output a pie chart is shown from main color of image to help user to decide what colors are exist in image. after entering the input color in hex number, user can see the segments which contain that color OR any close color to that and how many proprortion each segment is from whole image.

I used threshhod to help algorithm(kmeans) to choose not onlu user color but also close color to that in each segment since due to illumination variant in general image which are not taken in isolate situation in case of illuminance, occlusion and etc. ,it help to segment better each area which contain that similar colors.

in next day i am going to use GMM instead of k-means and also some better automate threshold like OStu Threshholdng to improve that. other deep learning based methods like Unet and RCNN based can be used with pretrained weights of Coco, backbone of Resnet101 and etc. can be tested too in next days