NVR/How To Set Up Face Recognition

How To Set Up Face Recognition

Description

Today we will look at the steps involved in setting up Face Recognition.

Prerequisites

• Luminys recorder that supports Face Recognition
• Luminys Recorder that has AI by recorder and can add Face Recognition, or a camera that has face recognition built in.
• LumiViewer if you want to use it for monitoring.

Steps

1. We will be using the WebUI for this process however it can also be done through the local monitor. The first step would be to log into the recorder through the WebUI by accessing the IP address of the NVR. Once logged in the first place we want to go to is “Camera” along the top.

2. For this document we will be using a camera that has face detect built in, but it does not have face recognition built in, so we will need to use the AI by recorder function. Because of this, we need to make sure LumiSearch is disabled in the recorder as it will prevent the usage of the AI by recorder function. To do this, click on the AI Event option on the left to drop down and see additional options, then click intelligent mode and finally on the right click on “Recorder AI Mode”. Once there you want to make sure the LumiSearch Mode is disabled. Of note, when you disable it, it will restart the NVR.

3. Once you have confirmed LumiSearch is not on, we can move onto setting up our face detection which is required for face recognition as well as setting up face recognition. From within the same area that you were in for the last step, this time click on Face Detection on the left. Once there, on the right side choose the channel you want to use for face detection and face recognition. Once you have chosen your channel, the first thing you want to make sure is that the “Type” is set for Smart Recorder function, that way we are using AI by recorder. Next you can go ahead and enable the Face Detection option. If you want to set up any other functions in this section you can but it is not necessary. Once done make sure to hit apply at the bottom.

4. The next step is now to work on the face recognition side of things. Before we can use face recognition, we first need to create a database that it will reference. You will also need to make sure you have pictures of people to add to the database. Once you have your pictures ready, you can go ahead and click on the add button on the right.

5. After clicking add it will bring up the add option. You will leave the type as “Common Database” and then add a name for the database which will help describe what the database contains such as “Employees”. Once you give it a name click ok and it will add that newly created database to the list.

6. To the right of the newly created database, you will see a circle with an (i) in it, click on that to start adding people into your database.

7. After clicking on the (i) you will be brought to the database where you will add the people you want in the database. Click on “Personnel Registration which will bring you into the option to add a person.

8. On this page you will have different sections to fill in information. Hold off on doing that until after you have added a photo. Click on the hashed box on the left to choose a photo.

9. After clicking on the hashed box where the picture goes, you will get a windows dialogue box to locate your image. Once you have located your image, click on it and then click open.

10. After clicking open, you will be taken to a screen where you need to select the photo and then click ok. By default the photo will not be selected, you need to click on it and then click ok.

11. You will then be brought back to the registration page where you have the different sections to fill out information. Go ahead and fill in whatever information you want for the person and then click ok at the bottom.

12. You will then be brought back to the main database page. You can go ahead and add more users if you wish, or you can close out of the page and move onto the next steps. On my monitor the x to close this window was off page, so you may need to scroll to the right a little in your browser to see the x to close. (My image I already moved over to the right which is why the left looks cut off)

13. Now we will be back to the face database section. You will notice it now shows “Register No.” as a 1 since I added 1 person to the database, and then you will notice the status is currently “Not Armed” we will see that change later on when we finish setting everything up.

14. Now that we have our database set up, now we can move onto setting up the face recognition. Under the AI Event section, let’s click on “Face Recognition” on the left. Now from here, you will want to choose the same channel on the right that we set face detection on. Make sure the “Type” is set for Smart Recorder function, then go ahead and enable Face Recognition. Once you have done that, you will see the box in the bottom left of the face recognition section. You will want to click on the + icon within that box.

15. By clicking the + button it will bring up the face database selector. Add a check box to the face database you want to use, you can use multiple if you have more than one. After selecting your face database by adding the check click ok.

16. You will then be brought back to the face recognition page for the channel you were setting up. It is very important that you add a check box into the face database within that box in the bottom left this is what actually makes the database active, adding it alone into that box without putting a check in the box won’t make it active. You will also notice there is a similarity of 80% there by default. If you want to change how similar the face needs to be to detect you can click on the 80% and then type in a new number. Lastly, once you have a database and have this all set up, in the bottom of the options on the right, you will see a stranger alarm linkage option. If your database has all of the people added that should be allowed in the building, then you can toggle on the strange alarm and then what will happen is if the system detects a face of a person that doesn’t exist in the database it will tag that person as a stranger, this can be very helpful if this is a building where only certain people should be allowed. This is completely optional. Once you have chosen all of your settings, click apply in the bottom left.

17. Now that we have everything set up we can give it a test. Go to the Live View section of the NVR, then so that you can see the results, switch the live view mode to target mode as it will display any results immediately. You will see that on the left it has detected the person from the database itself and shows how similar the image is, then below that an example of a stranger appearing that is not in the database.

18. Next you can also use the Smart Search function to search historically for any face captures or face recognition captures. To do so click the search button along the top of the UI, then on the left expand the Smart Search option, once expanded click on face search. Then on the right you will see 2 options, face detection and face search. Face detection will just search for any faces detected and then face recognition will search for any recognized faces from the database. For this we will show the face recognition option. Choose the channel you want to search or you can choose all cameras, choose the date range you want to search and then you can select specific attributes or just search for all, lastly you can also choose the similarity. Once you have chosen you options click search.

19. Once you see the search results, you can then export the results if you want, you can select a result and play it back as well. Once done, you can click the little x in the top right to return to the search page.

20. Next we will move onto using LumiViewer with the NVR to see how it works. For this part we will not be discussing how to add the recorder as that is handled in another document, this will only go over the Face Detect/Face Recognition function within LumiViewer. The first thing you want to do in LumiViewer is go to the Event Config section so we can setup the face recognition event function.

21. Once in the event config, you will want to locate the NVR and expand it so you can see your cameras, then select the camera that you have face recognition set up on. Once you have done that, in the middle section you will see a list of different types of events. For this we will want to select the target recognition option. Now you will see a bunch of different options on the right, to ensure this function will work, the main thing you need to enable is the target recognition toggle at the top. Then so we can see it occur, we will also enable the details pop up. Make sure to save in the bottom right to save your changes.

22. Now we will run a test with the settings we have set. You don’t need to be live viewing the camera for this to work, you just need to have LumiViewer open. When you walk in front of the camera and it detects a face it recognizes, it will pop up an event window with details. You can then watch playback of the event and then process the event.

23. You can also do the same with the face detection, or if you turned on the strange function, that also has an option in LumiViewer to have it do the same thing. There are quite a few options to use such as sending an email, playing a sound through the LumiViewer client.

24.Lastly, lets learn how to search for results using LumiViewer. First thing you want to do is go to the home page of LumiViewer and from there click on Data Retrieval.

25. Once in the data retrieval section, you will want to choose face search on the left. From there you will want to choose your NVR from the device list, and then choose whichever channels you want to search within. Then below the device listing, we will want to choose our event type, this list has the options for face detection, target recognition as well as stranger. For this we will use the target recognition option. Once you have selected that option click search along the bottom and you will see any search results on the right. You can then click on any of the results to see more details about the event.