如何 save/crop 在 dlib python 中检测到人脸

Question

我想通过裁剪矩形将检测到的人脸保存在dlib中任何人都知道我该如何裁剪它。我第一次使用 dlib 并且有这么多问题。我也想运行 fisherface 算法检测到的面孔，但是当我将检测到的矩形传递给 predictor 时，它给我类型错误。在这个问题上我非常需要帮助。

import cv2, sys, numpy, os
import dlib
from skimage import io
import json
import uuid
import random
from datetime import datetime
from random import randint
#predictor_path = sys.argv[1]
fn_haar = 'haarcascade_frontalface_default.xml'
fn_dir = 'att_faces'
size = 4
detector = dlib.get_frontal_face_detector()
#predictor = dlib.shape_predictor(predictor_path)
options=dlib.get_frontal_face_detector()
options.num_threads = 4
options.be_verbose = True

win = dlib.image_window()

# Part 1: Create fisherRecognizer
print('Training...')

# Create a list of images and a list of corresponding names
(images, lables, names, id) = ([], [], {}, 0)

for (subdirs, dirs, files) in os.walk(fn_dir):
    for subdir in dirs:
        names[id] = subdir
        subjectpath = os.path.join(fn_dir, subdir)
        for filename in os.listdir(subjectpath):
            path = subjectpath + '/' + filename
            lable = id
            images.append(cv2.imread(path, 0))
            lables.append(int(lable))
        id += 1

(im_width, im_height) = (112, 92)

# Create a Numpy array from the two lists above
(images, lables) = [numpy.array(lis) for lis in [images, lables]]

# OpenCV trains a model from the images

model = cv2.createFisherFaceRecognizer(0,500)
model.train(images, lables)

haar_cascade = cv2.CascadeClassifier(fn_haar)
webcam = cv2.VideoCapture(0)
webcam.set(5,30)
while True:
    (rval, frame) = webcam.read()
    frame=cv2.flip(frame,1,0)
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    mini = cv2.resize(gray, (gray.shape[1] / size, gray.shape[0] / size))

    dets = detector(gray, 1)

    print "length", len(dets)

    print("Number of faces detected: {}".format(len(dets)))
    for i, d in enumerate(dets):
        print("Detection {}: Left: {} Top: {} Right: {} Bottom: {}".format(
            i, d.left(), d.top(), d.right(), d.bottom()))

    cv2.rectangle(gray, (d.left(), d.top()), (d.right(), d.bottom()), (0, 255, 0), 3)


    '''
        #Try to recognize the face
        prediction  = model.predict(dets)
        print "Recognition Prediction" ,prediction'''





    win.clear_overlay()
    win.set_image(gray)
    win.add_overlay(dets)

if (len(sys.argv[1:]) > 0):
    img = io.imread(sys.argv[1])
    dets, scores, idx = detector.run(img, 1, -1)
    for i, d in enumerate(dets):
        print("Detection {}, score: {}, face_type:{}".format(
            d, scores[i], idx[i]))

Answer 1

请使用最少工作的示例代码以更快地获得答案。

在检测到人脸之后 - 你就有了一个矩形。所以你可以裁剪图像并用opencv函数保存:

    img = cv2.imread("test.jpg")
    dets = detector.run(img, 1)
    for i, d in enumerate(dets):
        print("Detection {}, score: {}, face_type:{}".format(
            d, scores[i], idx[i]))
        crop = img[d.top():d.bottom(), d.left():d.right()]
        cv2.imwrite("cropped.jpg", crop)

Answer 2

应该是这样的：

crop_img = img_full[d.top():d.bottom(),d.left():d.right()]

Answer 3

很好，但它忽略了原始矩形部分位于图像之外的边缘情况 window。（是的，dlib 会发生这种情况。）

crop_img = img_full[max(0, d.top()): min(d.bottom(), image_height),
                    max(0, d.left()): min(d.right(), image_width)]

Answer 4

# Select one of the haarcascade files:
#   haarcascade_frontalface_alt.xml  
#   haarcascade_frontalface_alt2.xml
#   haarcascade_frontalface_alt_tree.xml
#   haarcascade_frontalface_default.xml
#   haarcascade_profileface.xml

我记得haarcascade_frontalface_alt.xml是最好的？

如何 save/crop 在 dlib python 中检测到人脸

how to save/crop detected faces in dlib python

python

opencv

face-detection

face-recognition

dlib