Featuring insights from executives of cynapse.ai:
 |
 Gadi Piran US COO |
 |
Douglas Miorandi SVP, Market Development, US |
Spotlight #5 – Natural Language Search: From Query to Verified Result
Searching through hours of video footage has traditionally meant manually scrubbing through timestamps and camera feeds. This clip looks at how natural language search changes that, turning unstructured footage into structured, searchable data.
- Natural language querying — users can search using everyday descriptions, such as a specific vehicle type or color, across footage without any technical expertise or specialized training
- Locate specific incidents and scenarios, scored by confidence — operators can narrow results down to specific attributes and scenarios, with each match classified by a confidence score
- Set user-defined events—operators can define specific attributes, violation, or anomaly to watch for, so the system flags matching events automatically
- Generate incident summary report — results can be turned directly into a structured incident summary report, rather than requiring someone to write one up manually after the fact
It’s a concrete look at how the platform makes footage genuinely searchable — turning raw video into a usable, queryable record rather than something only makes sense when reviewed manually, frame by frame.
Watch the full episode and explore the rest of the series here: TTC Podcast – Intelligent Transportation Systems
