绘画作品的艺术风格包含画像的色彩、纹理、笔触等等,往往很难准确定义。基于深度学习的画像风格迁移让机器自行学习艺术家的作品,使得最终输出的合成图像呈现出与艺术家作品一致的风格。本演示可以把艺术家的风格迁移到摄像头捕捉的图像中,让平淡的图像呈现出不一样的艺术气息。(The artistic style of a painting, which includes colors, textures, strokes, etc., is often difficult to define accurately. Deep learning-based portrait style transfer allows the machine to learn the artist's work by itself, so that the final output composite image presents a consistent style with the artist's work. This demo can transfer the artist's style to the image captured by the camera, so that the dull image presents a different artistic atmosphere.)