- Regular meetings with the development team to provide suggestions on algorithm development and evaluation.
- Survey the latest state-of-the-art CNN and RNN based algorithms for facial anti-spoofing making use of multiple modal data inputs, provide suggestions on algorithms with a trade-off between accuracy and complexity
- Retrain selected models based on published databases
- Consultancy on industry direction, product strategy and academic research scope. Help formulate and focus research and development direction on facial anti-spoofing system. Provide innovative ideas for patent application.
- Milestone 1:
- Problem Discussion and approaches: To leverage existing expertise and know-how in CNN and RNN based algorithms for facial anti-spoofing
- Further review: Helping ASTRI team to review the latest state‐of‐the-art CNN and RNN based algorithms for facial anti-spoofing making use of multiple modal data inputs, provide suggestions on algorithms with a trade-off between accuracy and complexity
- Milestone 2:
- Open Lecture: One open technical lecture to be given on technical trend for the latest state-of-the-art CNN and RNN based algorithms for facial anti-spoofing.
- Technical Meetings: To join three meetings with the development team, helping to solve the technical problems encountered by the development team, which include retraining selected models, checking algorithms results, finding out the cause of the problem and improvement solutions
- Milestone 3:
- Open Lecture: One open technical lecture to provide innovative ideas on CNN and RNN anti-spoofing model for patent application
- Technical Meetings: To join three meetings with the development team, to discuss possible innovation ideas for patent applications which may include patent landscape search and claim definitions
- Ph.D. degree in EE, Computer Science or relevant disciplines.
- 15+ years of academic research experience in face recognition, biometric system security, data privacy and medical informatics is highly preferred. Candidate with less experiences may also be considered.
- Good understanding of facial anti-spoofing algorithm development and optimization is highly preferred.
- Strong publication in world top conferences/journals in image processing, biometrics, pattern recognition and artificial intelligence is highly preferred.
- Broad knowledge and connection in the community-keep abreast of the state of art of the latest progress, other industry players approach and academic R&D directions is highly preferred
Interested candidates please send application (quoting Ref. No.) with detailed resume, current and expected salary to Talent Acquisition via email to [email protected]
Only short-listed candidates will be notified. ASTRI reserves the right not to fill the position.
ASTRI is an Equal Opportunities Employer. Personal data provided by job applicants will be used exclusively for recruitment only.