mirror of https://github.com/opencv/opencv.git
Open Source Computer Vision Library
https://opencv.org/
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
94 lines
3.3 KiB
94 lines
3.3 KiB
import numpy as np |
|
import cv2, cv |
|
|
|
def detect(img, cascade): |
|
rects = cascade.detectMultiScale(img, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30)) |
|
if len(rects) == 0: |
|
return [] |
|
rects[:,2:] += rects[:,:2] |
|
return rects |
|
|
|
def detect_turned(img, cascade): |
|
img = cv2.cvtColor(img, cv.CV_BGR2GRAY) |
|
img = cv2.equalizeHist(img) |
|
|
|
img_t = cv2.transpose(img) |
|
img_cw = cv2.flip(img_t, 1) |
|
img_ccw = cv2.flip(img_t, 0) |
|
r = detect(img, cascade) |
|
r_cw = detect(img_cw, cascade) |
|
r_ccw = detect(img_ccw, cascade) |
|
|
|
h, w = img.shape[:2] |
|
rects = [] |
|
rects += [(x1, y1, x2, y2, 1, 0) for x1, y1, x2, y2 in r] |
|
rects += [(y1, h-x1-1, y2, h-x2-1, 0, -1) for x1, y1, x2, y2 in r_cw] |
|
rects += [(w-y1-1, x1, w-y2-1, x2, 0, 1) for x1, y1, x2, y2 in r_ccw] |
|
return rects |
|
|
|
def process_image(fn, cascade, extract_faces=True): |
|
img = cv2.imread(fn) |
|
h, w = img.shape[:2] |
|
scale = max(h, w) / 512.0 |
|
small = cv2.resize(img, (int(w/scale), int(h/scale)), interpolation=cv2.INTER_AREA) |
|
rects = detect_turned(small, cascade) |
|
|
|
for i, (x1, y1, x2, y2, vx, vy) in enumerate(rects): |
|
cv2.rectangle(small, (x1, y1), (x2, y2), (0, 255, 0)) |
|
cv2.circle(small, (x1, y1), 2, (0, 0, 255), -1) |
|
cv2.putText(small, str(i), ((x1+x2)/2, (y1+y2)/2), cv2.FONT_HERSHEY_PLAIN, 1.0, (0, 255, 0)) |
|
|
|
rects = np.float32(rects).reshape(-1,6) |
|
rects[:,:4] = np.around(rects[:,:4]*scale) |
|
|
|
faces = [] |
|
if extract_faces: |
|
path, name, ext = splitfn(fn) |
|
face_sz = 256 |
|
for i, r in enumerate(rects): |
|
p1, p2, u = r.reshape(3, 2) |
|
v = np.float32( [-u[1], u[0]] ) |
|
w = np.abs(p2-p1).max() |
|
fscale = w / face_sz |
|
p0 = 0.5*(p1+p2 - w*(u+v)) |
|
M = np.float32([u*fscale, v*fscale, p0]).T |
|
face = cv2.warpAffine(img, M, (face_sz, face_sz), flags=cv2.WARP_INVERSE_MAP | cv2.INTER_AREA) |
|
faces.append(face) |
|
|
|
return small, rects, faces |
|
|
|
|
|
|
|
if __name__ == '__main__': |
|
import sys |
|
import getopt |
|
from glob import glob |
|
from common import splitfn, image_extensions |
|
|
|
args, img_args = getopt.getopt(sys.argv[1:], '', ['cascade=', 'outdir=']) |
|
args = dict(args) |
|
cascade_fn = args.get('--cascade', "../../data/haarcascades/haarcascade_frontalface_alt.xml") |
|
outdir = args.get('--outdir') |
|
|
|
img_list = [] |
|
if len(img_args) == 0: |
|
img_list = ['../cpp/lena.jpg'] |
|
else: |
|
for mask in img_args: |
|
img_list.extend(glob(mask)) |
|
img_list = [fn for fn in img_list if splitfn(fn)[-1].lower() in image_extensions] |
|
|
|
cascade = cv2.CascadeClassifier(cascade_fn) |
|
|
|
for i, fn in enumerate(img_list): |
|
print '%d / %d %s' % (i+1, len(img_list), fn), |
|
vis, rects, faces = process_image(fn, cascade) |
|
if len(faces) > 0 and outdir is not None: |
|
path, name, ext = splitfn(fn) |
|
cv2.imwrite('%s/%s_all.bmp' % (outdir, name), vis) |
|
for face_i, face in enumerate(faces): |
|
cv2.imwrite('%s/%s_obj%02d.bmp' % (outdir, name, face_i), face) |
|
print ' - %d object(s) found' % len(faces) |
|
cv2.imshow('img', vis) |
|
cv2.waitKey(50) |
|
cv2.waitKey()
|
|
|