Better late than never...ramaz16 wrote: ↑Tue Apr 06, 2021 4:52 am BVR Trimming
It would be nice to trim and save an existing BVR section as a separate BVR without re-encoding. Allows to save plenty of disk space in home setup, if done daily to keep only small video sections of interest. When required, some of these sections can be converted later to MP4, possibly in bulk on a more powerful PC. However, conversion might make them inadmissible into evidence in some jurisdictions, where video evidence is often treated by courts as "hearsay", especially if its not the original footage.
Why Trim feature wasn't added for all years BI exists? Its obvious not everyone wants or able to keep tons of useless recordings, or re-encode select sections daily for no particular reason, but only because its impossible to trim the BVR and preserve its overlay timestamp.
BI 5 Feature suggestion thread
Re: BI 5 Feature suggestion thread
Re: BI 5 Feature suggestion thread
The ability to see what recordings a user has watched.
We currently have the ability to see how many frames a user has watched and the duration of the video under the Connections tab in Status window but having the ability to see exactly what videos a user reviewed would be very useful for auditing users to make sure they're watching the right things.
We currently have the ability to see how many frames a user has watched and the duration of the video under the Connections tab in Status window but having the ability to see exactly what videos a user reviewed would be very useful for auditing users to make sure they're watching the right things.
Re: BI 5 Feature suggestion thread
And who makes that determination?
Using Blue Iris to spy on its own users is certainly a novel use case. Not sure I'd call it a 'feature'.
Seems like key and screen loggers in the respective workstations of concern would be a more effective technique.
Re: BI 5 Feature suggestion thread
Choose the alert label for an object in an AI configuration by selecting the video frame with the largest bounding box. This approach is typically more effective because larger objects within an image are generally recognized with higher accuracy.
This approach addresses my issue where a fox at a distance is mistakenly identified as a cat by the AI, often with a confidence level greater than 80%. As the fox moves closer to the camera, it is recognized as a fox with a confidence level exceeding 95%. Setting the confidence threshold above 95% could result in missing some fox detections.
This approach addresses my issue where a fox at a distance is mistakenly identified as a cat by the AI, often with a confidence level greater than 80%. As the fox moves closer to the camera, it is recognized as a fox with a confidence level exceeding 95%. Setting the confidence threshold above 95% could result in missing some fox detections.