As discussed in the Solution Overview section, EPIC iO DeepInsights supports many use cases well suited to extracting insights from data, including video feeds and telemetry. This section presents the findings of our selected use case validations that highlight most of the essential features of DeepInsights.  Our use case validation used video loops provided by EPIC iO and the video analytics supported by DeepInsights. The use cases selected for testing depended on the availability of non-proprietary sample videos that capture typical scenes in a public transportation environment. Our short segments of test video were constantly looped to simulate a live feed from each of the nine viewing positions reasonably representative of a large urban commuter bus.

This section describes how we used each of the tools available from the DeepInsights solution to create end-to-end workflows as they would be implemented by an organization providing public bus transportation services. The focus was on functionality, ease of use, and practicality of the insights produced.

This portal is used to configure the mapping between video streams from the on-vehicle cameras to the video analytics engine that performs AI inferencing to create events. The portal also displays scenes from the bus on screen so users can configure one or more Regions of Interest (ROI) for each specific camera view. Rules are then defined for each ROI to detect events. We will describe each of these functions in detail.