Software General Issues

1. Where is the software setup bar?

This server uses Linux operating system so the software setup bar is in the top column of the operation system instead of in the utility interface of each program. When you move cursor to top column of the desktop, the setup bar will appear in the upper-left corner as shown in below graph. The specific parameters of each program can be config in this setup bar.

2. How to move the setup bar into program utility interface

Linux supports to display the setup bar in the program utility interface. Please do following actions to achieve this function.

a. Click on the system icon in the left-upper tool bar, type “appearance” in and click the icon of appearance as shown in below graph.

b. Choose the tab “Behavior” and then check the option of “in the window' s title bar” as shown in the following graph

After the setup is complete, the setup bar will be shown in the utility interface as below graph.

3. How many video input sources can be supported by this program？

This server provides four kinds of application analysis programs and these programs are mainly for video content analysis. The program supports three kinds of video sources input including local Video, network Video from RTSP Protocol camera and USB camera. User can choose different input source in setup bar.

4. How to input the video source by network video？

Network video means the video is captured from the camera which supports RTSP Protocol and transfer to server through internet. These cameras are commonly used in security surveillance system. Different cameras have their corresponding IP address that server can get relative video stream and other information from each camera then do video analysis. The operation process is as following:

a. Select the video source as net video which is shown in the following figure.

b. When you select successful, the following configuration dialog box pops out:

c. The meaning of each parameters are listed as below：

admin:hk123456 means Username: password

10.30.37.200:554 means Network address of IP Camera: Network Port

0/main/av_stream means this information of video stream transmission path

Please see the IP camera user manual or contact your IP camera provider to get related Information if you don’t know which value should be set.

d. The procedure of switching to different network video source

Please remember that close the current camera first then repeat the step.b to set to another network camera.

Face detection and recognition program

1. The difference between Face Detection and recognition

Face detection means the program will only detect and describe where is a person's face in a picture or video without recognizing who he is. When the program detects there is a human face in the picture, there will be a green box with the detected face inside. The example is shown as below graph:

Face recognition means the program will compare the detected face with existing enrolled user database and identify who he/she is. If the face is recognized successfully, his/her name will be shown beside the green box as below graph。

Both these two functions are all implemented in our face recognition program.

2. Introduction of face recognition operation process

First of all, collect and record the face of all people in need to build the enrolled user database. There is an icon “add person” shown in the utility. Press the “add person” icon then the camera will start to detect the face in the video source. Once the face is detected and pop out the green box, please press the “add person” icon, then program will start to record the face inside the green box and pop out another information icon that you can fill the name and relative information for this face. After all the needed faces are recorded, please press the “register all data” icon to let program store all the recorded faces and relative data. After the enrolled user database is build, when any face is detected in the video steam, program will determine if the detected face match to the face in the database or not and provide the relative name or show unknown.

3. What should be paid attention to when face recording?

The most important thing of face recording is how detail and precisely the face can be recorded. It is highly recommended that when doing the face recording, please input as much different view angels as possible. For example, at least in upright frontal position then rotate -45, -30, 30 and 45 degrees for face side view recording.

4. How to modify the threshold of face recognition

Setting value of threshold can be modified in setting bar. Click Config->Algorithm config, as shown in the following figure

The higher the threshold, the higher the recognition precision and vice versa

After the setup is complete, the new parameters will be activated after disable and re-enable the video input stream.

5. How many faces can be recognition in the same time?

How many faces can be recognized in the same flame can be controlled by the number set. Click Config->Algorithm Config, as shown in the following figure

Please input how many faces you want to detect in the same frame as shown in below:

After the setup is complete, the new parameters will be activated after disable and re-enable the video input stream.

Vehicle detection and classification procedure

1. Is there any limit of vehicle detection and classification procedure?

The vehicle detection and classification program in this server is implemented by the depth neural network model algorithm. This algorithm is designed for specific scenario with excellent performance but can’t applied to all kinds of customized scenario. If you have special requirement, welcome to contact us directly and we will be glad to provide the customized service.

2. How many kinds of objects can be identified in our vehicle detection and classification program?

There are fourteen kinds of objects can be identified and will be tagged as Minibus、Special_vehicle、SUV、Motercycling、Bus、Sedan_car、Smalltruck、MPV、Pedestrian、Bigtruck、Motorcycle、Minivan、Bike and Lorry.