[tool] [tutorial] Free AI Person Detection for Blue Iris

I'm trying to implement the older command line Docker version but I'm having trouble figuring out the right syntax for the config.txt file. I have three cams and three separate alerts, but when I list the 3 trigger URL's in the config file all three alerts fire. How do I get the config set so the 3 cams have 3 independent alerts?

Using the deprecated versions v0.6 and below you need to have the program three times, one per camera. That's one great advance of the new gui version, it can handle all cameras. Instead of using the deprecated version, I'd recommend to use the latest gui version, it is compatible with Deepquest AI in docker aswell in case oyu want to use the Deepquest Docker instead of the new installer.
 
  • Like
Reactions: sansnil
Is it possible to reset the stats? I see they are stored in the individual camera file but they seem to be admin locked.

The camera config files in ./camera/ usually should not be blocked, so you can simply close aitool.exe, edit the stats and restart aitool.exe. If for some reason you can't alter the camera ocnfig file, copy it's content, delete it and recreate the file.
 
Bit of a strange one here. see attached. The vehicle is masked out (150 op) but AI still detects it. ??

Please pm me the image, the mask file and the log.txt so that I do some tests.
Today I noticed that the software sometimes does not load the mask, I haven't yet figured why. To check if that is the reason, please open the log.txt file and check if one of the following errors occurs:

Code:
 ERROR: The resolution of the mask './camera/[cam name].png' does not equal the resolution of the processed image. Skipping privacy mask feature.
Code:
 ERROR while loading the mask file ./cameras/[cam name].png.

In v1.54 I modified the log a bit to get to the bottom of this issue, so a log of v1.54 would be most useful.


Please
 
Last edited:
Using the deprecated versions v0.6 and below you need to have the program three times, one per camera. That's one great advance of the new gui version, it can handle all cameras. Instead of using the deprecated version, I'd recommend to use the latest gui version, it is compatible with Deepquest AI in docker aswell in case oyu want to use the Deepquest Docker instead of the new installer.

Hi, thanks for your reply. I was mostly trying to solve for getting DQAI to auto-start on my Windows server but went down the wrong path in thinking I also had to use the command line version of AI Tool so I appreciate you sorting me out there. I like the GUI version of DQAI but it seems like it's been 2 years and they still haven't added a way to make it a service or otherwise autostart so I'm guessing it's not coming anytime soon which is too bad.

Thanks for your work here and providing a way to use a local instance of AI with Blue Iris!!! The directions are detailed and it's really cool to get less false positive alerts!
 
I'm having a lot of issues with the AI randomly stopping. I have log everything enabled, any ideas? If I restart, it will work for a little while then fail again.

Code:
[21.07.2019, 23:47:35]: Starting analysis of D:\BlueIris\AI_ID/DoorwayAI.20190721_234735073.jpg
[21.07.2019, 23:47:35]: 1. uploading image to DeepQuestAI Server ...
[21.07.2019, 23:47:35]: 2. Waiting for results ...
[21.07.2019, 23:49:15]: ERROR: Processing the image D:\BlueIris\AI_ID/DoorwayAI.20190721_234735073.jpg failed. Enabling 'Log everything' might give more information.
 
In my environment I don't have an issue with AITool stopping... its been running fine (processing a lot of alerts) for about 10 hours now...
 
In my environment I don't have an issue with AITool stopping... its been running fine (processing a lot of alerts) for about 10 hours now...

For me initially it was fine, recently it's started playing up. It would randomly stop processing the images.

Environment at the moment: Windows 1903
DeepStack running on Windows
AI Tool 1.55
 
Same version of Window and AITool as me.

I am running DeeStack on separate virtual server (linux and docker) since I had this exact same issue.

After BI5 was released I decided to install it on a newly build Win10Pro - it was the 1903 ISO I downloaded for that. Manually migrated my BI settings across but I had similar issues as you running a local DeepStack.
Never got sorted. I posted in DeepStack forums - very slow progress and they are still looking at the issue trying to fix it.

My issues was not that it was working for a limited period of time - it was not working at all.
 
Same version of Window and AITool as me.

I am running DeeStack on separate virtual server (linux and docker) since I had this exact same issue.

After BI5 was released I decided to install it on a newly build Win10Pro - it was the 1903 ISO I downloaded for that. Manually migrated my BI settings across but I had similar issues as you running a local DeepStack.
Never got sorted. I posted in DeepStack forums - very slow progress and they are still looking at the issue trying to fix it.

My issues was not that it was working for a limited period of time - it was not working at all.

Might need to give Docker version another go, but I am done with Ubuntu. Bothers me so much, when the debian distros are much faster in my environment. Blue Iris recently has been riddled with memory leaks, and weird issues too. Hopefully that's the end of it though, since the latest updates were meant to sort it.
 
I'm having a lot of issues with the AI randomly stopping. I have log everything enabled, any ideas? If I restart, it will work for a little while then fail again.

Code:
[21.07.2019, 23:47:35]: Starting analysis of D:\BlueIris\AI_ID/DoorwayAI.20190721_234735073.jpg
[21.07.2019, 23:47:35]: 1. uploading image to DeepQuestAI Server ...
[21.07.2019, 23:47:35]: 2. Waiting for results ...
[21.07.2019, 23:49:15]: ERROR: Processing the image D:\BlueIris\AI_ID/DoorwayAI.20190721_234735073.jpg failed. Enabling 'Log everything' might give more information.

I'm seeing this too.
 
I suggest to check your Windows event viewer and under Windows logs (Application and/or System) see what errors you can see.
I posted on page 1 my AITool error and in here are the errors I see the event viewer - Windows10 1903 - DeepstackAI crashes
Check and see if your errors (if you have any) are the same as mine.
Might also help if you post in that thread to make this issue more visible and "encourage" the vendor to actively work on it :)
 
I'm having a lot of issues with the AI randomly stopping. I have log everything enabled, any ideas? If I restart, it will work for a little while then fail again.

Code:
[21.07.2019, 23:47:35]: Starting analysis of D:\BlueIris\AI_ID/DoorwayAI.20190721_234735073.jpg
[21.07.2019, 23:47:35]: 1. uploading image to DeepQuestAI Server ...
[21.07.2019, 23:47:35]: 2. Waiting for results ...
[21.07.2019, 23:49:15]: ERROR: Processing the image D:\BlueIris\AI_ID/DoorwayAI.20190721_234735073.jpg failed. Enabling 'Log everything' might give more information.

This exact same thing is happening to me. Running DQAI in a separate docker and it seems to be responding just fine, but my log is only supplying the same information you have listed here.

Is there a way you can paste a way to test the DQAI? Like a URL POST or something? I have log everything enabled and it's not giving me any further detail, indicating it's a problem with DQAI. I can access my DQAI interface via web browser just fine. A separate "test" button would be infinitely helpful.

Having said that, I'm encouraged that I will eventually get this working... this is great work!

As an edit to this message, I’m seeing that DeepStack is detecting and analyzing my image, its just erroring between posting the results and your program reading it.

From docker:
[GIN] 2019/07/23 - 23:02:02 | 200 | 9.967µs | 192.168.1.5 | POST /v1/vision/detection
[GIN] 2019/07/23 - 23:02:12 | 200 | 10.598µs | 192.168.1.5 | POST /v1/vision/detection

Do you have any plans on open sourcing your tool? it would greatly help in troubleshooting and further advancing it.
 
Last edited:
Is your Windows firewall blocking it?
Try and disable it and see if it works then
 
It is not possible to autostart DeepQuestAI at the moment. I anticipate that the great developers at DQAI will change this in the future :). Meanwhile, if you need autostart you might want to use the Docker version of DQAI, detailed setup instructions can be found in the spoiler of the deprecated versions in the first post. In my experience this comes at the price of slower analyses.

I emailed them and got the following reply;

'Thanks for reaching out. We will provide support for this in new releases and keep you updated once it is available.'

:)