Beijing Seetatech Technology Co., Ltd (Seetatech) provides technology, talent, and knowledge services in the area of face recognition, computer vision, and machine learning, functioning as a research centre for business customers. We aim at bringing the self-taught and evolvable research, development and innovation ability of AI to every enterprise, enhancing our customers’ competence in the era of Artificial Intelligence.
Seetatech originally came from the Vision Information Processing and Learning (VIPL) team in the Institute of Computing Technology, Chinese Academy of Sciences (ICT, CAS), focusing on face recognition and computer vision. The Chairman/CTO of Seetatech, Prof. SHAN Shiguang, has been working on computer vision, especially face recognition, for about 20 years. Prof. SHAN received Annual Young Scientist Award of CCF 2015, co-founded the Vision and Learning Seminar (VALSE) and was appointed in the first VALSE as the president of the steering committee and the board chairman. The CEO of Seetatech, Dr. LIU Xin, previously worked in the VIPL team. Dr. LIU received two 1st place awards and one 2nd place award in world contests of computer vision, including the 1st Place Award in Alibaba Large-Scale Image Search Challenge. Dr. LIU also founded the Deep Learning Class (DLClass) Wechat official account.
Since 1997, our team has been working on computer vision techniques, such as face recognition, in the Institute of Computing Technology, Chinese Academy of Sciences. We have kept a group of more than 10 people doing research and development on related subjects since 2000. Currently, we have over 40 members in the VIPL team (including postgraduate and doctoral students), and more than 20 members in the face recognition sector. In the past ten years, our VIPL team has completed or achieved:
- More than 200 papers (60+ on CCF A-Class journals and conferences)
- More than 10,000 citations by peers (Google Scholar statistics)
- Huawei Excellent Collaboration Achievements Award (2017)
- 1st Place Award in ICPR2016 ChaLearn Gesture Challenge (2016)
- National Natural Science Awards, Second Class (2015)
- 1st Place Award in Alibaba Large-Scale Image Search Challenge (2015)
- 1st Place Award in IEEE FG 2015 Video Person Recognition Evaluation (2015)
- 1st Place Award in ChaLearn Cultural Event Recognition (2015)
- 2nd Place Award in ChaLearn Apparent Age Estimation (2015)
- 1st Place Award in ACM IMCI 2014 Emotion Recognition in The Wild Challenge (2014)
- CVPR2008 Best Student Poster Award Runner-Up Award (2008)
- National Science and Technology Progress Award, Second Class (2005)
Sr. Technical AdvisorProf. GAO Wen,CAE Member, Professor at Peking University
Technical AdvisorDr. YU Kai,Founder of Horizon Robotics, previously at Baidu IDL
Technical AdvisorDr. HUA Gang,Principal Researcher at Microsoft Research Asia
Sr. Research AdvisorProf. CHEN Xilin,IEEE Fellow, Associate Director of ICT, CAS
CTOProf. SHAN Shiguang,Executive Deputy Director of Key Laboratory of Intelligent Information Processing, ICT, CAS
CEODr. LIU Xin,Two 1st place awards and one 2nd place award in world CV contests
Co-FounderDr. ZHANG Jie,Papers accepted by CVPR, ICCV, and ECCV
Co-FounderDr. YUAN Jirui,Sr. R&D Director, PhD from TUT, Finland
Co-FounderGUO Danqing,Developed Internet products with DAU of 10M+
In the past 20 years, our VIPL team has provided technology services for partners including ISVISION, Huawei, Baidu, China Mobile, Samsung, Qualcomm, Intel, OMRON, NEC, Panasonic. Our products and solutions can be seen everywhere, such as Shanghai Expo 2010, passport check in China, face recognition for Public Security Department in over ten Chinese provinces, and many smart mobile phones around China.
Seetatech’s technologies are developed and refined by our VIPL team in the past 20 years, with outstanding features below.
Seeta Face Detection
Seeta Face Detection can accurately locate the human face in the image. Funnel-shape all-posture human face detection technique developed by VIPL can pick out all kinds of faces in complicated environments, including side faces, emotions, and even partially-occluded faces. The accuracy has reached high standard in the world, with FDDB recall rate over 90% (100 false positives).
Seeta Facial Landmark Positioning
Seeta Facial Landmark Positioning can precisely find out the position of more than 80 facial landmarks in the image such as eyes, noses, eyebrows and mouths in real time. The CFAN facial landmark positioning technique developed by VIPL features high response speed, and has reached high standard in worldwide evaluation under IBUG 300-W and AFLW datasets.
Seeta Face Recognition
Seeta Face Recognition is based on deep learning, and reaches an accuracy of 99.5% on average in LFW evaluation. Moreover, it was a key reason that we won the 1st Place Award in IEEE FG 2015 Video Person Recognition Evaluation. Based on our human face data (10M+ face pictures) collected during years, we can provide accurate and fast face recognition technology for scenarios like ID check, face verification and blacklist surveillance.
Seeta Face Attribute Analysis
Seeta Face Attribute Analysis can judge the gender, age, emotion, colour, glasses, and hat of the person in the image. Relying on this technique, we won the 1st Place Award in ACM IMCI 2014 Emotion Recognition in The Wild Challenge (2014) and the 2nd Place Award in ChaLearn Apparent Age Estimation (2015).
Seeta General Vision Recognition
With the support of VIPL team, we have rich research experiences and the capability to realize various functions, such as detecting general objects, recognizing and classify images, partitioning the image, and generating the title of an image automatically.
Seeta Deep Learning
The Seetatech team is one of the earliest teams studying deep learning in China. We have published dozens of papers related to deep learning, and own rich research and development experiences in this realm, which leads us to a knowledge and technology provider in the market.
SeetaFace is a free and open-source face recognition engine, providing a suite of basic and automatic face recognition software system. SeetaFace includes face detection module “SeetaFace Detection”, facial landmark positioning module “SeetaFace Alignment”, and facial characteristics acquisition and comparison module “SeetaFace Identification”. SeetaFace is coded using C++ and does not rely on any third-party library. SeetaFace is under the 2-Clause BSD License and can be used for free in the academia and industry.
Download Link: https://github.com/seetaface/SeetaFaceEngine