图像&视频编辑工具箱MMEditing使用示例：图像抠图(matting)

      MMEditing的介绍及安装参考：https://blog.csdn.net/fengbingchun/article/details/126331541，这里给出图像抠图的测试代码，论文：《Indices Matter: Learning to Index for Deep Image Matting》：

      (1).下载模型(checkpoint)：

def download_checkpoint(path, name, url):
	if os.path.isfile(path+name) == False:
		print("checkpoint(model) file does not exist, now download ...")
		subprocess.run(["wget", "-P", path, url])
 
path = "../../data/model/"
checkpoint = "indexnet_mobv2_1x16_78k_comp1k_SAD-45.6_20200618_173817-26dd258d.pth"
url = "https://download.openmmlab.com/mmediting/mattors/indexnet/indexnet_mobv2_1x16_78k_comp1k_SAD-45.6_20200618_173817-26dd258d.pth"
download_checkpoint(path, checkpoint, url)

      (2).根据配置文件和checkpoint文件构建模型：

config = "../../src/mmediting//configs/mattors/indexnet/indexnet_mobv2_1x16_78k_comp1k.py"
model = init_model(config, path+checkpoint, device)

      (3).准备测试图像：

image_path = "../../src/mmediting/tests/data/"
image_name = "merged/GT05.jpg"
trimap_name = "trimap/GT05.png"

      每组需要2张，一张是待抠图的彩色图像；一张是三元图(trimap)，如下图所示：源图来自于MMEditing

       (4).进行推理抠图：

result = matting_inference(model, image, trimap) * 255

      (5).显示执行结果及保存图像：

print(f"result shape: {result.shape}; max value: {np.max(result)}") # result shape: (552, 800); max value: 255.0
cv2.imwrite("../../data/result_matting_indexnet.jpg", result)
cv2.imshow("show", result)
cv2.waitKey(0)

      结果图如下所示：

      GitHub： https://github.com/fengbingchun/PyTorch_Test