To configure a VMD:

1. The Video Motion Detection tool can receive tracking metadata generated by the Motion Mask. The metadata are recorded into the database by default. To disable, select No in the Record Object Tracking (1).

   
   

2. If a camera supports multistreaming, select the stream for which detection is needed (2). Selecting a low-quality video stream allows reducing the load on the Server.
3. To reduce false alarms rate from a fish-eye camera, you have to position it properly (4).  For other devices, this parameter is not valid.

4. If necessary, enable the video stream grooming (5). In this case, only the I-frames will be decoded.

   This setting applies to all codecs. If a codec has keyframes and p-frames, the keyframe is decoded no more often than every 500 milliseconds. For the MJPEG codec, each frame is considered to be I frame. This feature reduces the Server load but, as can be expected, negatively impacts the quality of detection. This setting should be activated on "blind" Servers (Servers that do not display video) on which it is necessary to perform detection.

   
   

   Period and Decode key frames parameters are correlated. If no local Clients are connected to a Server, the following rules are set for remote Clients: If the interval between consecutive I-frames exceeds the value specified in the Period field, the detection tool will process I-frames. If the I-frame frequency is lower than the value specified in the Period field, the detection tool will use the set value. If at least one local Client connects to the Server, the detection tool is forced to use the set value. After the local Client disconnects, the indicated rules become valid again.

   
   

5. Select a processing resource for decoding video streams (6). When you select a GPU, a stand-alone graphics card takes priority (when decoding with NVidia NVDEC chips). If there's no appropriate GPU, the decoding will use the Intel Quick Sync Video technology. Otherwise, CPU resources will be used for decoding.

6. The VMD analyzes differences between the current frame and the static background (10, 11). Depending on a particular scene, we recommend the following sensitivity values for contrast and size:

   1. Maximum sensitivity (street scenes where target objects are smaller):

      1. Sensitivity: contrast = 16.

      2. Sensitivity: size = 10.

   2. Medium sensitivity (default values for generic scenes):

      1. Sensitivity: contrast = 12.

      2. Sensitivity: size = 9.

   3. Low sensitivity (indoor cameras with an average distance to object of ca. 4m):

      1. Sensitivity: contrast = 8.

      2. Sensitivity: size = 8.

         For your convenience with setting sensitivity value, in the preview window you can see the Motion Mask. To disable it, select No in the Motion Mask (7) field.

         
         If there is motion, but it does not exceed the threshold value (because of the detection sensitivity), the mask cells are colored green. If motion triggers VMD, the cells turn red.

7. To get tracked objects and their parameters (percentage of the FoV width/height, color) displayed in the Preview window, select Yes in the Object Tracking field (8). 

8. In the Period field (9) enter the time in milliseconds. This the time before  the next video frame is processed. This value should be in the range [0, 65535]. If the value is 0, each frame is processed.

9. By default, VMD (video motion detection) covers the entire FoV. In the FoV, you can set privacy masks - closed areas, inside of which you want no detection.
   Privacy masks are created similar to scene analysis configuration  (see Setting General Zones for Scene Analytics). 

10. Click Apply.

VMD configuration is now complete.