Face Recognition in Enterprise Video Conferencing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video conferencing, there is a need for improved participant identification due to time constraints and the challenge of remembering names, especially when multiple participants are co-located in the same room, which existing technologies do not adequately address.
Innovation Solution
A system with a processor and storage that accesses a video conference invite to fetch participant photographs and locations, performs face recognition on co-located participants using enterprise personnel database images, and displays their names and live video images, with the ability to expand matching if unrecognized participants are found, and includes access to social network photographs for enhanced identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If face recognition is performed on all participants without location filtering, then recognition accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system segments the face recognition task by first identifying co-located participants through location data filtering, then performing face recognition only on that subset. This divides the processing workload into two stages: location-based grouping and targeted recognition, reducing overall processing time while maintaining accuracy for relevant participants.
Solution Approach 2:
The system applies location-based filtering to identify which participants are co-located in the same room, then focuses face recognition resources only on those local participants rather than processing all participants uniformly. This localizes the recognition effort to the most relevant targets, improving efficiency.
2Productivity
If the system filters participants by location before face recognition, then processing efficiency improves, but the system complexity increases
Solution Approach 1:
The location filtering mechanism serves multiple functions: it identifies co-located participants for recognition, determines which participants should be displayed together, and filters the database queries. This multi-functional approach reduces the need for separate processing stages and simplifies the overall system architecture.
Solution Approach 2:
The system performs location-based filtering and participant grouping as a preliminary action before executing face recognition. By pre-identifying which participants are co-located and should be recognized together, the system prepares the data structure in advance, reducing the complexity of the recognition process itself.
3Reliability
If the system uses only enterprise personnel database photographs, then privacy and security are maintained, but recognition coverage is limited
Solution Approach 1:
The system uses the enterprise personnel database as an intermediary layer that bridges internal employee identification with external recognition needs. By filtering and matching against this controlled database first, the system maintains security while still achieving comprehensive recognition coverage for all invited participants through the expandable matching mechanism.
Solution Approach 2:
The system initially performs recognition using only the enterprise personnel database (partial action), and only expands to include additional sources like social media photographs if necessary. This approach maintains security by default while providing versatility when needed, avoiding the need to continuously access external sources.
4Measurement precision
If formal introductions are conducted for all participants, then participant identification is accurate, but meeting time is consumed
Solution Approach 1:
The face recognition system performs self-service identification of participants automatically, eliminating the need for manual introductions. The system independently identifies who is present, matches them against the invite list, and provides name displays without requiring time-consuming verbal introductions or name tags.
Solution Approach 2:
The system replaces the mechanical process of manual introductions and name tags with automated face recognition technology. By using image processing, database matching, and automatic display, the system achieves participant identification without the time consumption of traditional human-based introduction protocols.
Data Source
AI summary
A video conference invite is accessed and names and/or account addresses in the invite are used to obtain information of people listed on the invite from a database. The information can include template photographs of the people and locations of the people. During a video conference that is subject of the invite, when it is determined that a video feed of the video conference indicates that plural participants are together in a single room, images of faces in the video feed are efficiently matched against only template photographs of people in the database indicated by the database as being associated with the room.