CAMShift stands for Continuously Adaptive Mean Shift. It is the basis for the face-tracking algorithm in OpenCV. It combines the basic Mean Shift algorithm with an adaptive region-sizing step. The kernel is a simple step function applied to a skin-probability map. The skin probability of each image pixel is based on color using a method called histogram backprojection. Color is represented as Hue from the HSV color model.
Face Tracking with CAMShift
import cv2 import cv2.cv as cv def camshift_tracking(img1, img2, bb): hsv = cv2.cvtColor(img1, cv.CV_BGR2HSV) mask = cv2.inRange(hsv, np.array((0., 60., 32.)), np.array((180., 255., 255.))) x0, y0, w, h = bb x1 = x0 + w -1 y1 = y0 + h -1 hsv_roi = hsv[y0:y1, x0:x1] mask_roi = mask[y0:y1, x0:x1] hist = cv2.calcHist( [hsv_roi], , mask_roi, , [0, 180] ) cv2.normalize(hist, hist, 0, 255, cv2.NORM_MINMAX); hist_flat = hist.reshape(-1) prob = cv2.calcBackProject([hsv,cv2.cvtColor(img2, cv.CV_BGR2HSV)], , hist_flat, [0, 180], 1) prob &= mask term_crit = ( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ) new_ellipse, track_window = cv2.CamShift(prob, bb, term_crit) return track_window def face_track(): cap = cv2.VideoCapture(0) img = cap.read() bb = (125,125,200,100) # get bounding box from some method while True: try: img1 = cap.read() bb = camshift(img1, img, bb) img = img1 #draw bounding box on img1 imshow("CAMShift",img1) except KeyboardInterrupt: break
from SimpleCV import Image, Camera cam = Camera() img = cam.getImage() time.sleep(0.4) # Move your face in some direction img1 = cam.getImage() bb = (100,200,100,100) # get Bounding Box from some method new_bb = img1.track(img,"camshift",bb) # do this in loop.
It seems that CAMShift is quite efficient and fast. I’ll add Lucas Kanede Tracker, Median Flow Tracker, Forward-Backward Error Tracker soon. And to top it all, I will add OpenTLD to SimpleCV.
P.S. Down with throat infection. Waiting for SimpleCV t-shirt and my Mid-Term evaluation.