如何從這種圖像中刪除背景？

Question

我想刪除此圖像的背景以僅獲取此人。 我有數千張這樣的圖像，基本上，一個人和一個有點發白的背景。

我所做的是使用像 canny 邊緣檢測器或 sobel 過濾器（來自skimage庫）這樣的邊緣檢測器。 然后我認為可以做的是，將邊緣內的像素變白，而將邊緣內的像素變黑。 之后，可以對原始圖像進行蒙版以僅獲取人的照片。

然而，使用canny邊緣檢測器很難得到一個封閉的邊界。 使用 Sobel 過濾器的結果還不錯，但是我不知道如何從那里開始。

編輯：

是否也可以去除右手和裙子之間以及頭發之間的背景？

Answer 1

以下代碼應該可以幫助您入門。 您可能希望使用程序頂部的參數來微調您的提取：

import cv2
import numpy as np

#== Parameters =======================================================================
BLUR = 21
CANNY_THRESH_1 = 10
CANNY_THRESH_2 = 200
MASK_DILATE_ITER = 10
MASK_ERODE_ITER = 10
MASK_COLOR = (0.0,0.0,1.0) # In BGR format


#== Processing =======================================================================

#-- Read image -----------------------------------------------------------------------
img = cv2.imread('C:/Temp/person.jpg')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

#-- Edge detection -------------------------------------------------------------------
edges = cv2.Canny(gray, CANNY_THRESH_1, CANNY_THRESH_2)
edges = cv2.dilate(edges, None)
edges = cv2.erode(edges, None)

#-- Find contours in edges, sort by area ---------------------------------------------
contour_info = []
_, contours, _ = cv2.findContours(edges, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)
# Previously, for a previous version of cv2, this line was: 
#  contours, _ = cv2.findContours(edges, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)
# Thanks to notes from commenters, I've updated the code but left this note
for c in contours:
    contour_info.append((
        c,
        cv2.isContourConvex(c),
        cv2.contourArea(c),
    ))
contour_info = sorted(contour_info, key=lambda c: c[2], reverse=True)
max_contour = contour_info[0]

#-- Create empty mask, draw filled polygon on it corresponding to largest contour ----
# Mask is black, polygon is white
mask = np.zeros(edges.shape)
cv2.fillConvexPoly(mask, max_contour[0], (255))

#-- Smooth mask, then blur it --------------------------------------------------------
mask = cv2.dilate(mask, None, iterations=MASK_DILATE_ITER)
mask = cv2.erode(mask, None, iterations=MASK_ERODE_ITER)
mask = cv2.GaussianBlur(mask, (BLUR, BLUR), 0)
mask_stack = np.dstack([mask]*3)    # Create 3-channel alpha mask

#-- Blend masked img into MASK_COLOR background --------------------------------------
mask_stack  = mask_stack.astype('float32') / 255.0          # Use float matrices, 
img         = img.astype('float32') / 255.0                 #  for easy blending

masked = (mask_stack * img) + ((1-mask_stack) * MASK_COLOR) # Blend
masked = (masked * 255).astype('uint8')                     # Convert back to 8-bit 

cv2.imshow('img', masked)                                   # Display
cv2.waitKey()

#cv2.imwrite('C:/Temp/person-masked.jpg', masked)           # Save

輸出：

Answer 2

如果您希望不使用紅色填充背景而是使其透明，您可以在解決方案中添加以下幾行：

# split image into channels
c_red, c_green, c_blue = cv2.split(img)

# merge with mask got on one of a previous steps
img_a = cv2.merge((c_red, c_green, c_blue, mask.astype('float32') / 255.0))

# show on screen (optional in jupiter)
%matplotlib inline
plt.imshow(img_a)
plt.show()

# save to disk
cv2.imwrite('girl_1.png', img_a*255)

# or the same using plt
plt.imsave('girl_2.png', img_a)

如果您願意，可以調整一些 png 壓縮參數以使文件更小。

下圖為白色背景。 或黑色 - http://imgur.com/a/4NwmH

Answer 3

作為替代方案，您可以使用這樣的神經網絡： CRFRNN 。

它給出了這樣的結果：

Answer 4

vs2017 的工作示例。
設置紅色背景但保存藍色..
還添加了透明示例。

如何去除女孩的身體，只留下照片中的裙子？ 有任何想法嗎？

# == https://stackoverflow.com/questions/29313667/how-do-i-remove-the-background-from-this-kind-of-image

import cv2
import numpy as np
from matplotlib import pyplot as plt

#== Parameters =======================================================================
BLUR = 21
CANNY_THRESH_1 = 10
CANNY_THRESH_2 = 200
MASK_DILATE_ITER = 10
MASK_ERODE_ITER = 10
MASK_COLOR = (0.0,0.0,1.0) # In BGR format


#== Processing =======================================================================

#-- Read image -----------------------------------------------------------------------
img = cv2.imread('img/SYxmp.jpg')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

#-- Edge detection -------------------------------------------------------------------
edges = cv2.Canny(gray, CANNY_THRESH_1, CANNY_THRESH_2)
edges = cv2.dilate(edges, None)
edges = cv2.erode(edges, None)

#-- Find contours in edges, sort by area ---------------------------------------------
contour_info = []
_, contours, _ = cv2.findContours(edges, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)
for c in contours:
    contour_info.append((
        c,
        cv2.isContourConvex(c),
        cv2.contourArea(c),
    ))
contour_info = sorted(contour_info, key=lambda c: c[2], reverse=True)
max_contour = contour_info[0]

#-- Create empty mask, draw filled polygon on it corresponding to largest contour ----
# Mask is black, polygon is white
mask = np.zeros(edges.shape)
cv2.fillConvexPoly(mask, max_contour[0], (255))



#-- Smooth mask, then blur it --------------------------------------------------------
mask = cv2.dilate(mask, None, iterations=MASK_DILATE_ITER)
mask = cv2.erode(mask, None, iterations=MASK_ERODE_ITER)
mask = cv2.GaussianBlur(mask, (BLUR, BLUR), 0)

mask_stack = np.dstack([mask]*3)    # Create 3-channel alpha mask

#-- Blend masked img into MASK_COLOR background --------------------------------------
mask_stack  = mask_stack.astype('float32') / 255.0          # Use float matrices, 
img         = img.astype('float32') / 255.0                 #  for easy blending

masked = (mask_stack * img) + ((1-mask_stack) * MASK_COLOR) # Blend
masked = (masked * 255).astype('uint8')                     # Convert back to 8-bit 

plt.imsave('img/girl_blue.png', masked)
# split image into channels
c_red, c_green, c_blue = cv2.split(img)

# merge with mask got on one of a previous steps
img_a = cv2.merge((c_red, c_green, c_blue, mask.astype('float32') / 255.0))

# show on screen (optional in jupiter)
#%matplotlib inline
plt.imshow(img_a)
plt.show()

# save to disk
cv2.imwrite('img/girl_1.png', img_a*255)

# or the same using plt
plt.imsave('img/girl_2.png', img_a)

cv2.imshow('img', masked)                                   # Displays red, saves blue

cv2.waitKey()

Answer 5

根據@jedwards 的回答，與 opencv4 一起使用時，會出現此錯誤：

Traceback (most recent call last):
  File "save.py", line 26, in <module>
    _, contours, _ = cv2.findContours(edges, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)
ValueError: not enough values to unpack (expected 3, got 2)

函數cv2.findContours()已更改為僅返回輪廓和層次結構

你應該改成這樣：

contours, _ = cv2.findContours(edges, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)

Answer 6

• 在獲得不完整的邊緣（如您所見）后，您可以運行閉合形態（一系列膨脹和侵蝕）（必須根據需要/邊緣狀態設置大小和迭代）。

• 現在假設您在主體周圍有一個恆定的邊緣，使用任何類型的填充算法 (blob) 組合邊緣對象外的所有點，然后取負數，為您提供對象內部的蒙版.