Global Filters
A rich selection of granular filters can be used to include and display only objects matching a range of characteristics on the VIEWER tab.
Refer to the following table for all available filtering options.
Filter applied | Case objects included |
Source | Objects originating from specific video synopses (if no sources are selected, objects from all sources will be displayed). By default, all sources are selected. Files that were uploaded will appear in a directory named FILES. |
Time Range | Objects matching specific time ranges. NoteOnly the selected sources are taken into account for the Time Range filter. |
Class | Objects matching the following classes: People: Man, Woman, Child Two-Wheeled Vehicles: Bicycle, Motorcycle Other Vehicles: Car, Pickup, Van, Truck, Bus, Train, Airplane, Boat Illumination Changes: Lights On, Lights Off (see also Illumination Changes) Animals Note that:
|
Person Attributes | Objects with the following attributes: Upper Wear: Long Sleeves, Short Sleeves, Colors* Lower Wear: Long, Short, Colors* Hat: No Hat, Hat – Hoods are considered hats. Face Mask: No Mask, Mask – The Face mask detector operates only on faces that are 1-star quality or above (objects that appear when you click the Show Faces option in the Face Recognition filter). Bag: No Bag, Backpack, Hand Held – Hand Held may also bring up any item held in a person’s hand, such as laptops and cell phones. You can fine tune the tolerance of the attributes by using the Tolerance field. See also Filter Tolerance Adjustment. You can fine tune the colors of the lower and upper wear by using the Shade and Coverage fields. See also Color Tolerance in Color Tolerance. *Note that:
|
Vehicle Make & Model | Vehicles matching a specific make and model. See also Vehicle Make and Model Recognition. |
Color | Objects matching any combination of brown, red, orange, yellow, green, lime, cyan, purple, pink, white, gray and black. When searching for people using colors in order to get the best possible results, it’s recommended to use the upper and lower colors in person attributes instead of the Color filter because the person attribute colors are optimized for people. See also Color Tolerance in Color Tolerance. |
Size | Objects based on their actual (real-life) size. You can select a size between 0-65 feet (0-20 meters). |
Speed | Objects based on their actual speed. You can select a speed between 0-155 mi/hr (0-250 km/hr). |
Dwell | Objects having dwelled for a user-specified period or longer in a scene. |
Direction | Objects having travelled in a specified direction. |
Proximity | Objects appearing below or above a specified distance. See also Proximity. |
Appearance Similarity | Objects with similar attributes, either people (People Similarity) or vehicles (Vehicle Similarity). See also Appearance Similarity. |
Face Recognition | Objects that have a detected face. See also Face Recognition. |
License Plate Recognition | Objects that have a detected license plate number. See also License Plate Recognition. |
BriefCam's classification method
BriefCam uses a two-tier classification approach. First, BriefCam classifies the class category (e.g. vehicle). For this top level, BriefCam has a 96% accuracy or above. For sub-classes the filtering is more detailed, and the accuracy might not be the same as the top-level classes (e.g. specific vehicle type).
In some cases, the fine-grain classification of sub-classes is not attempted. For example, if you have an object with a very low resolution, then it might be possible to detect that it’s a 2-wheeled vehicle, but the resolution is not high enough to determine if it’s a bicycle or a motorcycle.