Recently, I published a paper on image synthesis. Given a semantic layout (e.g., locations of cars and buildings), our cascaded refinement network can synthesize a high-resolution image that conforms to the input layout. The synthesized image can be thought as the visual imagination of machines.
我最近发布了机器生成图像的项目。我们的算法可以按照指定的布置图生成逼真的图像。生成的图片好比机器想象出来的画面。
Please watch our demo video. 请观看我们的介绍视频。
Project website: http://cqf.io/ImageSynthesis
Source code: http://github.com/CQFIO/PhotographicImageSynthesis
My GitHub project ranks #1 in all the Python repositories today! 我GitHub上的项目在所有Python项目里今天排第一名!
Photographic Image Synthesis with Cascaded Refinement Networks
Qifeng Chen and Vladlen Koltun
In ICCV 2017