基于mediapipe-selfie-segmentation的图像分割、背景切换解决方案

import cv2 import mediapipe as mp import numpy as np mp_drawing = mp.solutions.drawing_utils mp_selfie_segmentation = mp.solutions.selfie_segmentation # 图片人物抠图: IMAGE_FILES = ["face_example_2.jpg"] BG_COLOR = (0, 255, 0) # 背景颜色也可以使用其他的照片，要求与原照片尺寸一致 bg_image = cv2.imread('seasand3.jpg') height, width ,_ = bg_image.shape MASK_COLOR = (255, 255, 255) # mask图片颜色 with mp_selfie_segmentation.SelfieSegmentation(model_selection=0) as selfie_segmentation: for idx, file in enumerate(IMAGE_FILES): image = cv2.imread(file) image = cv2.resize(image,(width,height)) # 在处理之前需要转换图片到RGB颜色空间 image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) results = selfie_segmentation.process(image) # 在背景图像上绘制分割图 #为了改善边界周围的分割，可以考虑在 results.segmentation_mask进行双边过滤 # np.stack((results.segmentation_mask,) * 3, axis=-1) > 0.1最后参数越小，包括的边缘越多 condition = np.stack((results.segmentation_mask,) * 3, axis=-1) > 0.3 #生成纯色图像,白色的mask图纸 #fg_image = np.zeros(image.shape, dtype=np.uint8) #fg_image[:] = MASK_COLOR fg_image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR) if bg_image is None: # 背景为纯色 bg_image = np.zeros(image.shape, dtype=np.uint8) bg_image[:] = BG_COLOR output_image = np.where(condition, fg_image, bg_image) cv2.imshow('output_image',output_image) cv2.imwrite('output_image' + '.jpg', output_image) cv2.waitKey(0) #cv2.imwrite('selfie0.png', output_image)