I want in the image below to (by using Python):
1.) Find the contours of the bones (only the sides will do)
2.) Recognize and draw all the contours.
It could look something like this:
A better contour is even good. I am not entirely sure as to how I could tackle this,
The gradient of the image is:
A initial way of approaching this would be using canny edge detection using the right threshold values and then find the contours.
import cv2
# Load the image
img = cv2.imread("/home/tribta/Desktop/feet.png")
# Find the contours
imgray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
edges = cv2.Canny(img,60,200)
im2, contours, hierarchy = cv2.findContours(edges, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
hierarchy = hierarchy[0] # get the actual inner list of hierarchy descriptions
# For each contour, find the bounding rectangle and draw it
cv2.drawContours(img, contours, -1, (0,255,0), 3)
# Finally show the image
cv2.imshow('img',img)
cv2.waitKey(0)
cv2.destroyAllWindows()
Then you could add some biomedical processing criteria to distinguish the different contours and verify if it is really a bone.