On the morning of July 22, local time, the Kamehameha III Hall (Kamehameha III, named after the 19th century king of Hawaii, Kamehameha III) took a seat in the Hawaii Conference Center. The CVPR 2017 winning paper was formally announced here.
This year's CVPR has two best papers, two best paper nominations and one best student paper. Apple’s research on cyber Simng in December last year was one of the two best papers.
What other award-winning papers are there?
Best paper
Densely Connected Convolutional Networks
That is, "the densely connected convolutional network" by Gao Huang, Zhuang Liu, Laurens van der Maaten, Kilian Q. Weinberger. Several people are researchers at Cornell University, Tsinghua University and Facebook AI Institute (FAIR), and Kilian Q. Weinberge is a Computer Professor at Cornell University.
Thesis source code see here.
"Learning From Simulated and Unsupervised Images through Adversarial Training"
That is, "With help from training, learning from simulated, unsupervised images," by Ashish Shrivastava, Tomas Pfister, Oncel Tuzel, Joshua Susskind, Wenda Wang, Russell Webb.
This is the first AI paper released by Apple Inc., which pursues a secrecy culture, and marks the first step in opening up the door to open academic research. The paper was published in December last year and proposes a decomposed imGAN training method consisting of three parts (the simulator Simulator, the refiner Refiner, and a discriminator Discriminator).
Interestingly, some scholars initially questioned the gold content of the paper, saying that the significance of Apple's paper "test water" is far greater than the significance of the study itself. For details, please see the report of Lei Fengnet earlier this year.
Best Paper Nomination
Annotating Object Instance with a Polygon-RNN
That is, "Annotation of target instances with Polygon-RNN," by Lluis Castrejon, Kaustave Kundu, Raquel Urtasun, and Sanja Fidler are researchers at the University of Toronto. See the paper homepage here.
There is a child's shoes in the country that wrote a study note for the paper. To summarize its outline, see here.
YOLO 9000: Better, Faster, Stronger
The authors, Joseph Redmon and Ali Farhadi, are researchers at the University of Washington. In the title of this paper, "YOLO" is an abbreviation for "You Only Look Once" and is a real-time target detection system developed by the two authors. Its GitHub address here, 燡oseph Redmon introduces the study's personal home page poke here.
Best Student Paper Award
Computational Imaging on the Electric Grid
That is, "calculating imaging on electricity grids." The first author, Mark Sheinin, is a Ph.D. student at the Israeli Institute of Technology. The second and third authors, Yoav Y. Schechner and Kiriakos N. Kutulakos, are professors at the Israeli Institute of Technology and the University of Toronto respectively. This is a cross-cutting research in optics, computer vision, image processing, and electrical engineering. The authors developed a system for collecting information about the city's power grid by means of flashes that are not visible to the naked eye.
Award ceremony
燑br>
Type1 Tethered Cable,16A Type1 Tethered Cable,32A Type1 Tethered Cable,Type 1 Tethered Charging Cable
Yangzhou JERI New Energy Co., Ltd. , https://www.jrevcharging.com