New Features and Improvements

• Users should launch the Sentinel.AI application one time to complete the signup\training process.
• Option to guide the user in completing the signup process has been introduced. Instructions to guide the user in recording the video clip has been provided.
• Sentinel.AI Vision face detection & comparison accuracy has been improved significantly; users will now be requested to record a short video clip to perform face analytics.
• Option to preview the recording is also available. So, users can retake the video if the recorded video is not clear.
• Added functionality to capture and store OS version, Mac address and RAM capacity. This information is made available in Vision and Shield report agent detail view.
• All image analytics is performed on the local system.
• Sentinel.AI Vision captured images with are older than 24-hours are automatically purged from the local system to maintain the size of local database
• When Vision is not able to access the camera, a help page is provided to resolve the issue.
• Sentinel.AI health check has been introduced to ensure the application works as intended. When the health check detects a failure, it is reported back to the Vision report with an error code to help resolve the issue.
• If the camera is covered and all images captured or dark, then these images are not transmitted to the platform. Only no face detected anomaly information is sent for reporting.
• REST API token generation process has been re-engineered. Only one token will be generated for a 24-hour duration and repurposed for all subsequent interactions with the platform.
• Sentinel.AI will reload configuration from Sentinel.AI cloud once every 5 minutes.
• Bug fixes and performance improvements have been made.