What if the realistic portraits generated by GANs were turned into 3D versions?
It seems that there is a camera shooting straight at the portrait, and it can be displayed from different angles of the front, side, up, and down.
There really is a kind of smell that people want to come out of the painting.
Moreover, these effects are generated from static monoscopic images!
It can even make cartoon portraits three-dimensional.
This is the latest research from Shanghai Jiao Tong University and Huawei:CIPS-3D.
It is a GAN-based 3D perception generator that only uses raw single-view images without any upsampling to generate clear images with a resolution of 256×256.
And set a new record for 3D perception image synthesis, with an FID of only 6.97.
From 2010 to 2015, he was a research scientist at the Singapore Institute for Advanced Study at the University of Illinois at Urbana-Champaign.
He has published more than 550 articles, including 250+ IEEE TPAMI, IJCV, CVPR/ICCV/ECCV, NeurIPS and other top international journals and conferences.
Joined HUAWEI CLOUD in 2018, the main research directions are computer vision, natural language processing and voice interaction.
HuaweiXie LingxiPh.D., Shanghai Jiaotong UniversityNi BingbingProfessors also participated in the study.
Xie Lingxi, both undergraduate and Ph.D. degrees in computer science from Tsinghua University, specializing in computer vision and automatic machine learning. Currently a senior researcher at Huawei.
Ni Bingbing is currently a special researcher/tenured associate professor and doctoral supervisor of the Department of Electronics, Shanghai Jiaotong University.
He graduated from the Department of electronic Engineering of Shanghai Jiaotong University with a bachelor’s degree, and then went to the National University of Singapore to study for a Ph.D.
During his Ph.D., he worked successively in Microsoft Research Asia and Google’s US headquarters as an algorithm scientist.
From 2010 to 2015, he was a research scientist at the Singapore Institute for Advanced Study at the University of Illinois at Urbana-Champaign.
His research interests include computer vision and machine learning.
The Links: 3HAC7998-2 3HAC7343-1
