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

Whats the URL for deepstack? Use that in AITool
 
I'm running the docker version, so not sure if I'm helping ;)

I can watch the deepstack logs and see when an image is processed. Do you see any logs like this?

Code:
[GIN] 2019/08/22 - 12:25:13 | 200 |  1.888621888s |   192.168.XXX.YYY | POST     /v1/vision/detection
[GIN] 2019/08/22 - 12:26:05 | 200 |  1.555230341s |   192.168.XXX.YYY | POST     /v1/vision/detection
[GIN] 2019/08/22 - 12:26:13 | 200 |  1.505777399s |   192.168.XXX.YYY | POST     /v1/vision/detection
[GIN] 2019/08/22 - 12:31:45 | 200 |  2.461283118s |   192.168.XXX.YYY | POST     /v1/vision/detection
[GIN] 2019/08/22 - 12:31:50 | 200 |  2.764808968s |   192.168.XXX.YYY | POST     /v1/vision/detection
[GIN] 2019/08/22 - 12:31:53 | 200 |  1.553019084s |   192.168.XXX.YYY | POST     /v1/vision/detection

That's a good question. I searched the DepStack folders and didn't find anything that was dated recently and couldn't find a way to look at logs.


Whats the URL for deepstack? Use that in AITool

I've used the IP address (in my case 10.0.0.175 is the WAN side of my BI server) and "localhost" with port number with no success. The same IP goes in AIT.

Edit. Sorry if you were asking for the URL for DeepStack. It should be the IP address of the computer it is running on with the port.
 
Last edited:
Is 10.0.0.175 your WAN address or LAN address? Seems like a WAN address wouldn't work without some port forwarding unless the entire server is outside your DMZ.

If it's your WAN address, try your LAN IP:81 instead.
 
Woah, lots of replies.
I feel like I'm proficient in tech, so I believe I've covered my basics. I've confirmed that the DeepStack server and AITools/BlueIris are able to talk to each other and reach each others web page. I've gone through the standard disable firewall, defender, etc. Again, I can pull up the UI of BI and DeepStack on each machine without issues along with ICMP response.

Deepstack is running on port 81, AI Tools has IP only, no HTTP. Though, I did play and add http/https a few times to see if that helped the other night. Like Johnny2678 posted above, I see the deepstack receiving the data, but AI Tools pukes at the processing. Does anyone know of a way to enable a more verbose log for AITools or Deepstack?

Has anyone successfully got this setup on Windows 10 or Server 2016? I'm wondering if Windows is having an issue or there is a step missing in the instructions.

Originally when I set this up I set it up on a Windows 2016 box that BI is on. I added both AI Tool and DeepStack, but shortly removed Deepstack thinking there was some HTTP fighting. Didn't change my results, unfortunately. Is the original dev active here or should I navigate over to Deepstack's community?
 
Last edited:
I'm running BI5 and AITool on windows 10 machine. Both on the same machine. Also with Docker Desktop for Windows on the same machine.
 
Last edited:
I'm running BI5 and AITool on windows 10 machine. Both on the same machine. Also with Docker windows desktop on the same machine.

Docker may be the answer. I'm going to load this in docker and see if I get the same results. It seems everyone running the docker version is having success.

Thanks for the reply!
 
I did use the windows docker version successfully and that was ok but I changed over to the Docker Desktop for Windows version so I could run AITool as a service.
 
Last edited:
so localhost:81 should work in AITool. ??

Yes, it does for me. Localhost will only work while using the host machine. If you try to do it on a remote machine connected to the lan, use the IP address.

Perhaps antivirus is stopping it?

AV and firewall disabled. Still no go.

Is 10.0.0.175 your WAN address or LAN address? Seems like a WAN address wouldn't work without some port forwarding unless the entire server is outside your DMZ.

If it's your WAN address, try your LAN IP:81 instead.

That's my Wan IP. I do have two NIC's on the BI server to isolate the cameras so the WAN address is needed. I tried the camera lan with no success. All my testing is done on the machine to eliminate the internet.

It looks like I am going to have to give docker a try. I hope my limited computers skills can figure that out.

Thanks for all the help and

L8tr
d
 
Yes, it does for me. Localhost will only work while using the host machine. If you try to do it on a remote machine connected to the lan, use the IP address.



AV and firewall disabled. Still no go.



That's my Wan IP. I do have two NIC's on the BI server to isolate the cameras so the WAN address is needed. I tried the camera lan with no success. All my testing is done on the machine to eliminate the internet.

It looks like I am going to have to give docker a try. I hope my limited computers skills can figure that out.

Thanks for all the help and

L8tr
d

Interesting - can you share what you do? I have dual nic and didn't think this would work without some tweaking.

has anyone done a video on setup?
 
One NIC for WAN access and one NIC for Cameras using different subnet. Wan is 10.0.0.XXX and Cameras are 10.0.1.XXX. My machine is BI only and the controls LAN/WAN access. All cams are connected to switches and to one NIC. The WAN NIC is connected to my router. The cameras cannot access WAN due to subnet differences. BI uses the WAN IP address for the webserver. The rest of the house uses the WAN IP's for it's LAN(if that makes sense).

I hope that's what you were asking. Try searching for "subnet" and there should be tons of posts about it.

L8tr
d
 
One NIC for WAN access and one NIC for Cameras using different subnet. Wan is 10.0.0.XXX and Cameras are 10.0.1.XXX. My machine is BI only and the controls LAN/WAN access. All cams are connected to switches and to one NIC. The WAN NIC is connected to my router. The cameras cannot access WAN due to subnet differences. BI uses the WAN IP address for the webserver. The rest of the house uses the WAN IP's for it's LAN(if that makes sense).

I hope that's what you were asking. Try searching for "subnet" and there should be tons of posts about it.

L8tr
d

I am sorry. Yes I have the same setup. I was asking how to get this setup. But I don't have 10.0.0.xxx numbers I have 192.168.90.xxx
 
Code:
[GIN] 2019/08/22 - 12:25:13 | 200 | 1.888621888s | 192.168.XXX.YYY | POST /v1/vision/detection
[GIN] 2019/08/22 - 12:26:05 | 200 | 1.555230341s | 192.168.XXX.YYY | POST /v1/vision/detection
[GIN] 2019/08/22 - 12:26:13 | 200 | 1.505777399s | 192.168.XXX.YYY | POST /v1/vision/detection
[GIN] 2019/08/22 - 12:31:45 | 200 | 2.461283118s | 192.168.XXX.YYY | POST /v1/vision/detection
[GIN] 2019/08/22 - 12:31:50 | 200 | 2.764808968s | 192.168.XXX.YYY | POST /v1/vision/detection
[GIN] 2019/08/22 - 12:31:53 | 200 | 1.553019084s | 192.168.XXX.YYY | POST /v1/vision/detection

The below are the response times I am getting running DeepStack on a separate virtual server with docker.
I would like to be able to run DeepStack on the same Windows server as BlueIris but just as reported by others here I could never get to work. Never found a resolution for it.

Code:
[GIN] 2019/08/23 - 09:20:19 | 200 |  773.469306ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:20 | 200 |   815.93641ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:21 | 200 |  880.283939ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:22 | 200 |  894.129656ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:24 | 200 |  993.971403ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:25 | 200 |  969.609994ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:26 | 200 |  919.700132ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:28 | 200 |  911.294107ms |   192.168.0.xxx | POST     /v1/vision/detection
 
Last edited:
The below is the response times I am getting running DeepStack on a separate virtual server with docker.
I would like to be able to run DeepStack on the same Windows server as BlueIris but just as reported by others here I could never get to work. Never found a resolution for it.

Code:
[GIN] 2019/08/23 - 09:20:19 | 200 |  773.469306ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:20 | 200 |   815.93641ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:21 | 200 |  880.283939ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:22 | 200 |  894.129656ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:24 | 200 |  993.971403ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:25 | 200 |  969.609994ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:26 | 200 |  919.700132ms |   192.168.0.xxx | POST     /v1/vision/detection
[GIN] 2019/08/23 - 09:20:28 | 200 |  911.294107ms |   192.168.0.xxx | POST     /v1/vision/detection

I like your response times better.
 
Yeah I like them too.
But I would have liked to see if they improve by running DeepStack locally on the BI server (non docker).
But it wont work for me regardless of what I will try.
 
Last edited:
Yes, it does for me. Localhost will only work while using the host machine. If you try to do it on a remote machine connected to the lan, use the IP address.



AV and firewall disabled. Still no go.



That's my Wan IP. I do have two NIC's on the BI server to isolate the cameras so the WAN address is needed. I tried the camera lan with no success. All my testing is done on the machine to eliminate the internet.

It looks like I am going to have to give docker a try. I hope my limited computers skills can figure that out.

Thanks for all the help and

L8tr
d


Looks like I can't use Docker. I am using Win10 Home. I need to upgrade to pro someday. I am wondering if using Win10 Home is part of the problem too.

L8tr
d
 
Looks like I can't use Docker. I am using Win10 Home. I need to upgrade to pro someday. I am wondering if using Win10 Home is part of the problem too.

L8tr
d

In my opinion, all Windows 10 machines that turn into a surprise box every half year due to enforced upgrading are not optimal regarding reliability, not to mention the privacy concerns. If you're able to get hands on a license for the long term support LTSB / LTSC version of Windows 10 through your employer then you should definitely use it imho, way more stable, no junk-apps and full control over your system.