为什么我不能将我的图像放在眼睛检测上？

Question

我想在cv2 检测到我眼睛的位置放置一个透明图像。我已经完成了主要的两个步骤，现在我需要将它们结合起来。

例如，这里是带有image transparency working 的输出，这里是带有eye detection working 的输出。脚本和图片在下面，我不知道该怎么办。

图片

dot_transparent.png

image.jpg

app.py

​​>

import os
import numpy
import cv2
from PIL import Image
from os.path import join, dirname, realpath

def upload_files():
    #https://github.com/Itseez/opencv/blob/master/data/haarcascades/haarcascade_frontalface_default.xml
    face_cascade = cv2.CascadeClassifier('/Users/matt/Python/LazerEyes/haarcascade_eye.xml')

    #https://github.com/Itseez/opencv/blob/master/data/haarcascades/haarcascade_eye.xml
    eye_cascade = cv2.CascadeClassifier('/Users/matt/Python/LazerEyes/haarcascade_eye.xml')

    img = cv2.imread('new.png')
    dot = cv2.imread('dot_transparent.png', cv2.IMREAD_UNCHANGED)

    gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    gray_to_place = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    
    img_h, img_w = gray.shape
    img_to_place_h, img_to_place_w = gray_to_place.shape

    faces = face_cascade.detectMultiScale(gray, 1.3, 5)
    for (x,y,w,h) in faces:
        roi_gray = gray[y:y+h, x:x+w]
        roi_color = img[y:y+h, x:x+w]
        eyes = eye_cascade.detectMultiScale(roi_gray)
        for (ex,ey,ew,eh) in eyes:
            dot = cv2.resize(dot, (eh, ew))
            # Prepare pixel-wise alpha blending
            dot_alpha = dot[..., :3] / 255.0
            dot_alpha = numpy.repeat(dot_alpha[..., numpy.newaxis], 3, axis=2)
            dot = dot[..., :3]

            resized_img = cv2.resize(dot, (eh, ew), interpolation = cv2.INTER_AREA)
            resized_img_h, resized_img_w, _ = resized_img.shape

            #pointsOnFace = []
            #integersToAppend = eh
            #pointsOnFace.append(integersToAppend)
            #print(pointsOnFace)

            roi_color[ey:ey+resized_img_h, ex:ex+resized_img_w, :] = resized_img

    cv2.imwrite('out.png', img)

Answer 1

将my earlier answer 合并到给定的代码中（并最小化生成的代码），解决方案可能如下所示：

import cv2
import numpy as np

face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('haarcascade_eye.xml')

img = cv2.imread('new.jpg')
dot = cv2.imread('dot_transparent.png', cv2.IMREAD_UNCHANGED)

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

faces = face_cascade.detectMultiScale(gray, 1.3, 5)

for (x, y, w, h) in faces:

    eyes = eye_cascade.detectMultiScale(gray[y:y+h, x:x+w])

    for (ex, ey, ew, eh) in eyes:

        # Filter out small detections, if you're only want to have the eyes
        if ew < 100 or eh < 100:
            continue

        d = cv2.resize(dot.copy(), (eh, ew))
        d_alpha = d[..., 3] / 255.0
        d_alpha = np.repeat(d_alpha[..., np.newaxis], 3, axis=2)
        d = d[..., :3]

        img[y+ey:y+ey+eh, x+ex:x+ex+ew, :] = \
            img[y+ey:y+ey+eh, x+ex:x+ex+ew, :] * (1 - d_alpha) + d * d_alpha

cv2.imwrite('out.png', img)

这就是输出（我过滤掉了小的检测，这样只有实际的眼睛会被覆盖）：

可能仍需要微调确切的位置，但我认为这是来自 Haar 级联分类器本身的问题。

----------------------------------------
System information
----------------------------------------
Platform:      Windows-10-10.0.16299-SP0
Python:        3.9.1
NumPy:         1.20.1
OpenCV:        4.5.1
----------------------------------------