02
FINE-TUNE YOUR INSTALLATION ACCORDING TO YOUR NEEDS
Configuration
In the configure tab, you can edit different settings :
- You can rotate the Video Stream (useful if you can’t physically reach a camera that was installed upside down)
- The amount of CPU cores dedicated to Vidireports – This can improve the FPS, but we recommend using 2 Cores
Uploads
You can configure the data uploads, in the uploads tab. Once there, a quick overview gives information about the current session.
Under the overview section, the “Last Upload” section lets you transfer data in a click and see when the last transfer occurred.
The “Get detailed dashboards” section contains links to VidiCenter, where you can see all your data and create dashboards, monitor your network and much more!
In the upload parameters tab, once deployed by clicking on it. You can configure the Frequency section, which allows you to choose when to upload :
Under the “Uploads policy” you have the choice between :
- One upload a day with a given time,
- or every N minutes.
The “interval [min]” parameters only applys when you choose “Every N minutes” as the uploads policy to specify “N”
The “Upload time [HH:MM AM/PM]” parameters only applys when you choose “Once a day at a given time” as the uploads policy to specify at what time to upload.
The Randomization part give you the possibility to randomize more or less the upload time. In the “Before[min.]” you chose a given number of minutes to extend the randomization range before the regular upload time, and “After[min.]” to extend after the regular upload time. (For exemple if you choose an upload once a day at 12:00 PM with “Before[min.]” set to 10 and “After[min.]” set to 15, the upload will occure randomly between 11:50 AM and 12:15 PM”)
The “Upper Limits” part allows you to choose the size of the upload, thanks to two parameters. “Maximum “events per upload” to set the number of maximum events to allow per upload (set to 10 000 by default), and “Maximum upload size [kBytes]” to choose the size of the upload (set to 8192 by default)
Once finished, just click the save button and the modifications will be taken into account.
Masks
In the control center, you can define masks on the video feed. You have two kinds of masks : negative masks, that remove part of the video so that it does not get analyzed by VidiReports, and positive masks, that define precisely the zone where the analysis should happen (any pixel outside of that zone will not be analyzed)
You can choose to hide some parts of the video feed with negative masks. The main reasons to do that is :
- A false detection may occur often at a given zone on the video feed (if a TV is facing the camera for exemple, every person appearing on TV will be detected). Masking the zone will prevent VidiReports from analyzing any face within that zone.
- To reduce the size of the image to analyze: when you use masks, the pixels under the masks are not analyzed so the amount of pixels to analyze at each frame is lower allowing for more FPS. Typically, if the bottom part of the image never sees faces (only legs), you can mask it to not use computing power on those pixels that never show a face.
You can set up to one positive mask (a mask that restricts the analysis to a specific region in the image) and up to eight negative masks (masks that block parts of the image). To draw a mask simply click and drag your mouse over the image; by default you will obtain a negative mask. After drawing the mask you can:
- move the mask: place your cursor within the mask and drag (for the positive mask, place the cursor on the edge)
- resize the mask: place the cursor on the lower-right corner of the mask and drag
- convert the mask to a positive mask: select the mask and press the ‘i‘ key; only one positive mask is available at any time
- delete the mask: select the mask and press the ‘delete‘ key