Customized facial ID
6 months
2 FTE: Head of Computer Vision, ML Engineer
Project
K2's project delivers impeccable celebrity recognition, surpassing industry standards, and offers ongoing customization within the client's infrastructure.
Results
The project stands out with its high-quality feature space, achieving recognition results akin to industry giants.
The solution's scalability allows ongoing development according to the client's evolving needs. Deployed within the client's infrastructure, it remains independent of external providers, ensuring data integrity and seamless customization.
Media Press
In collaboration with the client, K2 Summit undertook a visionary project to develop an application for precise celebrity recognition. The objective was to align seamlessly with the client's preferences and integrate effortlessly into their existing system. An additional challenge included the client's requirement for continuous expansion, enabling non-technical users to effortlessly add new faces to the system.
Delivering sustainable value.
Deliverables
Impact
Skills
Massive Dataset: Processed over 500,000 faces, creating a dataset representing 20,000+ unique individuals, surpassing public datasets. Deep Learning Model: Implemented a cutting-edge deep learning model for face recognition, storing unique features in an Elasticsearch-based vector database. Scalable System: Developed a solution that can continually evolve, enabling easy integration of new faces by non-technical users. Quality Feature Space: Achieved recognition results comparable to industry giants, emphasizing the importance of a high-quality feature space.
The project stands out with its high-quality feature space, achieving recognition results akin to industry giants. The solution's scalability allows ongoing development according to the client's evolving needs. Deployed within the client's infrastructure, it remains independent of external providers, ensuring data integrity and seamless customization.
Face Recognition, Deep Learning, Elasticsearch, Customizable Solutions, Scalable System