Blue Iris Updates (64 bit)

Blue Iris Updates - Official Thread 6.0.7.1

What exactly is this: AI object tracker power-up?
The “AI Object Tracker Power-Up” in Blue Iris is a newer enhancement added in recent Blue Iris 6 updates that improves how the built-in AI follows and evaluates objects across multiple frames instead of analyzing just a single image.

Here’s what it actually does in practice:
  • Tracks detected objects frame-to-frame
  • Determines whether an object is moving or stationary
  • Helps suppress false alerts from static objects
  • Improves AI reliability for people, vehicles, and animals
  • Feeds tracking data into alert confirmation logic
Blue Iris describes it as an added intelligence layer where the “Object tracker is used to follow and analyze objects for motion (or lack thereof) through consecutive frames.”

A big benefit is filtering out things like:
  • Parked cars
  • Trash cans
  • Shadows/light shifts
  • Objects that appeared briefly and stopped moving
Instead of repeatedly alerting on those, the tracker can classify them as static and ignore them.

It also improves:
  • Better person/vehicle continuity
  • Cleaner alert images with tracked bounding boxes
  • Object counting and movement analysis
  • More accurate “real-time AI object analysis” for things like license plates and scene activity
This is different from the older AI flow in Blue Iris, which mostly worked like:
  1. Motion detected
  2. Snapshot sent to AI
  3. AI says “person” or “no person”
  4. Alert generated or canceled
The new tracker adds temporal analysis — meaning it understands how objects behave over time, not just in one frame.

You’ll mainly notice improvements in:
  • Reduced false positives
  • Better stationary object filtering
  • More stable tracking in busy scenes
  • Smarter alert
 
The “AI Object Tracker Power-Up” in Blue Iris is a newer enhancement added in recent Blue Iris 6 updates that improves how the built-in AI follows and evaluates objects across multiple frames instead of analyzing just a single image.

Here’s what it actually does in practice:
  • Tracks detected objects frame-to-frame
  • Determines whether an object is moving or stationary
  • Helps suppress false alerts from static objects
  • Improves AI reliability for people, vehicles, and animals
  • Feeds tracking data into alert confirmation logic
Blue Iris describes it as an added intelligence layer where the “Object tracker is used to follow and analyze objects for motion (or lack thereof) through consecutive frames.”

A big benefit is filtering out things like:
  • Parked cars
  • Trash cans
  • Shadows/light shifts
  • Objects that appeared briefly and stopped moving
Instead of repeatedly alerting on those, the tracker can classify them as static and ignore them.

It also improves:
  • Better person/vehicle continuity
  • Cleaner alert images with tracked bounding boxes
  • Object counting and movement analysis
  • More accurate “real-time AI object analysis” for things like license plates and scene activity
This is different from the older AI flow in Blue Iris, which mostly worked like:
  1. Motion detected
  2. Snapshot sent to AI
  3. AI says “person” or “no person”
  4. Alert generated or canceled
The new tracker adds temporal analysis — meaning it understands how objects behave over time, not just in one frame.

You’ll mainly notice improvements in:
  • Reduced false positives
  • Better stationary object filtering
  • More stable tracking in busy scenes
  • Smarter alert
Thank you for the great explanation.

Unfortunately, in my case this feature it is actually preventing objects to being detected, even if they should be (not static). I'm getting much fewer alerts and that's totally wrong. I don't know if I'm doing something wrong, , but also an explanation on which are the configurations to improve the detection would be much appreciated!
 
The “AI Object Tracker Power-Up” in Blue Iris is a newer enhancement added in recent Blue Iris 6 updates that improves how the built-in AI follows and evaluates objects across multiple frames instead of analyzing just a single image.

Here’s what it actually does in practice:
  • Tracks detected objects frame-to-frame
  • Determines whether an object is moving or stationary
  • Helps suppress false alerts from static objects
  • Improves AI reliability for people, vehicles, and animals
  • Feeds tracking data into alert confirmation logic
Blue Iris describes it as an added intelligence layer where the “Object tracker is used to follow and analyze objects for motion (or lack thereof) through consecutive frames.”

A big benefit is filtering out things like:
  • Parked cars
  • Trash cans
  • Shadows/light shifts
  • Objects that appeared briefly and stopped moving
Instead of repeatedly alerting on those, the tracker can classify them as static and ignore them.

It also improves:
  • Better person/vehicle continuity
  • Cleaner alert images with tracked bounding boxes
  • Object counting and movement analysis
  • More accurate “real-time AI object analysis” for things like license plates and scene activity
This is different from the older AI flow in Blue Iris, which mostly worked like:
  1. Motion detected
  2. Snapshot sent to AI
  3. AI says “person” or “no person”
  4. Alert generated or canceled
The new tracker adds temporal analysis — meaning it understands how objects behave over time, not just in one frame.

You’ll mainly notice improvements in:
  • Reduced false positives
  • Better stationary object filtering
  • More stable tracking in busy scenes
  • Smarter alert
Can I use it to track a trash can instead of ignore it?

I’d like to know when/if the trash can was taken out to the curb on time on trash day
 
Unfortunately, in my case this feature it is actually preventing objects to being detected, even if they should be (not static). I'm getting much fewer alerts and that's totally wrong. I don't know if I'm doing something wrong, , but also an explanation on which are the configurations to improve the detection would be much appreciated!
There are already several threads explaining BI configurations. I suggest you use the search option to find what you are looking for so this thread doesn't get hijacked any more than it already has.
 
Unfortunately, in my case this feature it is actually preventing objects to being detected, even if they should be (not static). I'm getting much fewer alerts and that's totally wrong. I don't know if I'm doing something wrong, , but also an explanation on which are the configurations to improve the detection would be much appreciated!

Same issue here.

I've noticed that if a car or person passes a camera's FOV and out of sight and then returns not long after then the second activation and AI confirmed alert is dropped. I've got two BI systems with one for test purposes and running these back to back (6.0.6.9 vs 6.0.5.7) with the same settings makes this very apparent.
 
Same issue here.

I've noticed that if a car or person passes a camera's FOV and out of sight and then returns not long after then the second activation and AI confirmed alert is dropped. I've got two BI systems with one for test purposes and running these back to back (6.0.6.9 vs 6.0.5.7) with the same settings makes this very apparent.
Interesting; I haven't seen that in such a direct manner, but you may want to send your results to the BI developer (Ken). With empirical evidence showing this kind of degradation of performance, he is usually willing to listen and make adjustments as needed.
 
Interesting; I haven't seen that in such a direct manner, but you may want to send your results to the BI developer (Ken). With empirical evidence showing this kind of degradation of performance, he is usually willing to listen and make adjustments as needed.

Already emailed Ken before posting.

There were a lot of issues with v5 and early v6 with considering static objects as moving in cases such as more than one in a camera's view (AI detection ordering could vary), slight variation with detected bounding rectangle etc, so I hope we're not going too far the other way now. It would be good if Ken could expose more settings around timings for static/moving or just document its logic but I appreciate he keeps evolving the code which could render yet more settings obsolete and the help file often lags enough as it is.
 
Not sure if this is the right place to ask for this but i have an enhancement request that i think many would find useful... I know I would. It would be very helpful to be able to save an action using the description and then be able to pick from a list of those saved actions to easily add to other cameras. Most of my cameras have many actions but are used often.
 
Not sure if this is the right place to ask for this but i have an enhancement request that i think many would find useful... I know I would. It would be very helpful to be able to save an action using the description and then be able to pick from a list of those saved actions to easily add to other cameras. Most of my cameras have many actions but are used often.
A Blue Iris support ticket is the only way feature requests will make it to the developer.
 
Blue Iris Settings > About tab > Get support button. Then there's an Email Support section near the bottom of that panel. Nicely hidden away right??
thanks for that explanation... well, just sent an email per the instructions... we shall see.