Organizational Structure and Industry and Academic Research Partnerships

Headquartered in Beijing, Megvii Research operates four research institutes in China. Led by Megvii's Chief Scientist, Dr. Sun Jian, the institutes are applying the company’s deep learning technology to advance AI research and application development. Megvii is also proactively building a partnership that combines industry and academic research in order to enhance its core R&D capabilities and cultivate emerging talent in the field of AI.

Research Areas

Deep Learning

Computer Vision

Visual Navigation and Control

Computational Photography

Machine Learning Systems

Robotics

Deep Learning

Computer Vision

Visual Navigation and Control

Computational Photography

Machine Learning Systems

Robotics

Megvii Facts & Figures

Megvii has won 49 world championships across several top global AI competitions, including a record-breaking three consecutive COCO Challenge titles

Megvii has been involved in setting more than 30 national and industry standards related to AI

Megvii has more than 1,300 business-related authorized patents

Megvii grants access to its 1400+ R&D staff to make use of Brain++

Brain++ has helped reduce algorithm production cost by 50%

Brain++ has helped improve AI application development efficiency by 80%

Megvii has won 49 world championships across several top global AI competitions, including a record-breaking three consecutive COCO Challenge titles

Megvii has been involved in setting more than 30 national and industry standards related to AI

Megvii has more than 1,300 business-related authorized patents

Megvii grants access to its 1400+ R&D staff to make use of Brain++

Brain++ has helped reduce algorithm production cost by 50%

Brain++ has helped improve AI application development efficiency by 80%

Publications

Publications 学术人物学术活动

旷视研究院ShuffleNet V2斩获VALSE年度杰出学生论文奖

4 月13 号正会第一天，VALSE 2019 于合肥如期而至，现场正式公布由评奖委员会终评确定的“VALSE年度杰出学生论文奖”，旷视研究院明星论文 ShuffleNet V2（ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design）从海选论文中“杀入”Short List，最终脱颖而出，拿下 “VALSE 2018 年度杰出学生论文奖”的殊荣。这与旷视研究院的产学研体系和旷视 Brain++ 的背景支撑密不可分。

计算机视觉VALSE深度学习

ECCV 2018 | 10篇论文+5项世界第一，记旷视科技ECCV之旅

当地时间 9 月 14 日，欧洲计算机视觉顶级学术会议 ECCV 2018 在德国慕尼黑圆满落幕。据悉，ECCV 2018 规模空前，有近 3200 人参加，接收论文 776 篇；另有 43 场 Workshops 和 11 场 Tutorials。旷视研究院在院长孙剑博士的带领下远赴盛会，用心用力，推动全球范围计算机视觉的技术交流与产品落地。

计算机视觉深度学习CVPRECCV

ECCV 2018 | 旷视包揽COCO+Mapillary四项世界第一，中国公司成最大赢家

当地时间 9 月 8 日，两年一度、为期一周的欧洲计算机视觉顶级学术会议 ECCV 2018 在德国慕尼黑拉开帷幕。据官方信息显示，ECCV 2018 共有 43 场 Workshops 和 11 场 Tutorials；大会投稿论文 2439 篇，其中接收论文 776 篇（31.8%）；注册参会人数近 3200 人，规模空前，超过 ECCV 2016（阿姆斯特丹）2 倍。旷视研究院在院长孙剑博士的带领下远赴盛会，用心用力，推动全球范围计算机视觉的技术交流与产品落地。

ECCVCOCOMapillary竞赛

<1 2 3 4 >

Telephone Consultation

Online ConsultationContact Online Customer Service for Product Details

Business Cooperation