ljw2k
Known around here
Yes I have created an account and have the activation Key but for some reason my PC is saying the site is unsafe and tried to get it to work on many browsers to allow access but just fails.
[tool] [tutorial] Free AI Person Detection for Blue Iris
- Thread starter GentlePumpkin
- Start date
pmcross
Pulling my weight
Are you getting the certificate warning when trying to access the DeepStack srver on your local machine or when trying to access DeepStack AI Server ? If from your local IP, then you can ignore the certificate warning as this is due to DeepStack using a self-signed SSL certificate.
ljw2k
Known around here
Yes I am getting that warning which stops me accessing their page where I need to activate it via the Key. I have tried everything but can't get on their page to activate,
GentlePumpkin
IPCT Contributor
- Joined
- Sep 4, 2017
- Messages
- 193
- Reaction score
- 322
news: The creators of Deepstack plan to make deepstack open source and entirely free, which is a very nice decision. I think they were unable to make the project comercially profitable, I guess they intend to make their living from consultation regarding DQAI. Edit: Just read that at least one of the makers now works at Microsoft, well deserved I would say! Anyways, this means, that we will be able to have a permanent, zero maintenance, private AI Backend + the ability to customize it and create custom detection APIs for free. Maybe it will be possible to teach the AI and thus continously enhance detections, that would be super-incredible!
So prepare yourself for some nice enhancements and features as soon as I find time to utilize the upcoming possibilities
cheerfully
GentlePumpkin
So prepare yourself for some nice enhancements and features as soon as I find time to utilize the upcoming possibilities
cheerfully
GentlePumpkin
Last edited:
ljw2k
Known around here
Can anyone help me how to register this key please,
GentlePumpkin
IPCT Contributor
- Joined
- Sep 4, 2017
- Messages
- 193
- Reaction score
- 322
maybe it's simply because of missing https?
You could try internet explorer (probably the only useful reason for existence of this browser besindes accessing malicious china ip cams
)
ljw2k
Known around here
HaHa ...tried that mate
GentlePumpkin
IPCT Contributor
- Joined
- Sep 4, 2017
- Messages
- 193
- Reaction score
- 322
So you can't get past the browser warning?
ljw2k
Known around here
That's correct
ljw2k
Known around here
It states site not secure in any browser I try for instance Chrome below 
GentlePumpkin
IPCT Contributor
- Joined
- Sep 4, 2017
- Messages
- 193
- Reaction score
- 322
created a key for you, here you go: f6e0aa20-c07f-4751-a71a-96bc9ba404ae
ljw2k
Known around here
I have a key already got that when I registered as below and not sure how or where to register it as when I start the server it states i need to register the key
ljw2k
Known around here
It states Visit Localhost to visit . I Hit Dev Center to register the key but that site won't open no matter which browser I use.
GentlePumpkin
IPCT Contributor
- Joined
- Sep 4, 2017
- Messages
- 193
- Reaction score
- 322
so you open localhost:81 or whatever port you chose and then you simply enter the key.
GentlePumpkin
IPCT Contributor
- Joined
- Sep 4, 2017
- Messages
- 193
- Reaction score
- 322
amazing idea, I always wanted to implement sth like that. Maybe even showing so relevant additional details when hovering over the rectangle. And btw thank you for your commit and contribution, I'll update the downloads here soon.
I think I never did it because that would need to rewrite some details of the image box. The image box I currently use can offer one overlay over the image - that's used for the rectangles now. So you 1. need to know when a mouse is above a specific rectangle and 2. eigther change the existing overlay and add the information (not a very clean solution but it will work) or use another image showing solution.
And without my fancy hovering idea, maybe just composing the overlay from the existing + single images containing the names of possibly detected objects (dirty solution regarding different screen sizes again) would do it.
Or any other, probably cleaner, implementation.
I think I never did it because that would need to rewrite some details of the image box. The image box I currently use can offer one overlay over the image - that's used for the rectangles now. So you 1. need to know when a mouse is above a specific rectangle and 2. eigther change the existing overlay and add the information (not a very clean solution but it will work) or use another image showing solution.
And without my fancy hovering idea, maybe just composing the overlay from the existing + single images containing the names of possibly detected objects (dirty solution regarding different screen sizes again) would do it.
Or any other, probably cleaner
, implementation.
Last edited:
I was going to draw text over the top left of the rectangle, but this has a few obvious issues (out of bounds text and scaling issues). I think I will start and try to implement it first and add onto it later if needed, let me know if you have any suggestions as I am new to this and appreciate any help I can get.
Im running DeepQuestAI with docker on a ubuntu container. The AI Tool shows me:
Deepstack AI Server logs:
Code:
[06.05.2020, 16:55:38]: ERROR: Processing the following image 'C:\BlueIris\aiinput/Gar1sd.20200506_165537420.jpg' failed. Can't reach DeepQuestAI Server at 192.168.1.97:5000.
[06.05.2020, 16:58:57]: ERROR: Processing the following image 'C:\BlueIris\aiinput/Terr1sd.20200506_165857461.jpg' failed. Failure in AI Tool processing the image.
[06.05.2020, 16:59:00]: ERROR: Processing the following image 'C:\BlueIris\aiinput/Ht1sd.20200506_165900095.jpg' failed. Failure in AI Tool processing the image.
Code:
root@deepstack-ai:~# docker run -e VISION-SCENE=True -v localstorage:/datastore -p 80:5000 deepquestai/deepstack
/v1/vision/scene
---------------------------------------
v1/vision/addmodel
---------------------------------------
v1/vision/listmodels
---------------------------------------
v1/vision/deletemodel
---------------------------------------
---------------------------------------
v1/backup
---------------------------------------
v1/restore
[GIN] 2020/05/06 - 14:53:32 | 403 | 62.932µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:53:32 | 403 | 20.514µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:53:35 | 403 | 28.752µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:53:35 | 403 | 38.22µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:03 | 403 | 36.487µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:07 | 403 | 24.05µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:18 | 403 | 40.339µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:18 | 403 | 17.595µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:23 | 403 | 27.691µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:23 | 403 | 20.268µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:34 | 403 | 22.565µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:55:37 | 403 | 23.506µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:58:57 | 403 | 30.685µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:58:57 | 403 | 20.633µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:59:00 | 403 | 21.714µs | 192.168.1.17 | POST /v1/vision/detection
[GIN] 2020/05/06 - 14:59:00 | 403 | 20.754µs | 192.168.1.17 | POST /v1/vision/detectionLooks like you may not have the detection API enabled, I have not used the docker container but usually you see /v1/vision/detection on startup. I also see that a HTTP 403 not found error is being sent instead of the normal 200 OK.
I see you are right. I installed DeepStack AI Server on Windows and it even wont work. At Windows Server 2019 I see /v1/vision/detection, but it even wont work.