Target group definition is nothing new – in fact, it is a well-established discipline for customer service, advertising, store layout, etc. Correct identification of the audience affects the planning, creation and implementation of all activities. Inaccurate systems or even manual identification can lead to errors.
- So far collected data included all people counts, but no information was captured on the proportion of male and female customers/visitors
- Unfiltered data fed into an analytics tool leads to obscure results and the conclusions drawn are based mostly on assumptions rather than facts
- Activities are planned by guesswork, using additional expensive systems or inaccurate manual identification of female/male target groups
The Xovis all-in-one solution
Xovis sensors now allow you to identify male and female target groups. Benefit from all existing functions and add value simply by installing the Gender Statistics* plugin.
- Sensors count all people, recognize whether a person is female or male and thus identify this basic target group
- AI firmware has been trained to visually distinguish people whether they are female or male. When the plugin is activated, the sensor or AI firmware mimics the human eye. Its focus lies on facial features, the overall upper body and movement patterns
- Male and female passers-by are visualized live in the WebUI for easy verification
- Data provides the proportion of female and male customers/visitors for further analytics
- No additional expensive system required
- AI-powered Xovis sensors running an AI-based people counting algorithm provide an embedded solution. Everything is computed on the sensor, with no need for external processing power.
*Statement regarding the term gender: Xovis respects and embraces all dimensions of diversity, including gender identity anywhere along or beyond the spectrum of gender expression. For technical reasons, the algorithm in the Xovis Gender Statistics feature recognizes only easily discernible, visual indications when determining whether a person is more likely to be female or male. A reliable detection of the biological sex of a person is neither possible nor intended. We intend no disrespect to the gender with which a person identifies. The counts are merely a statistical measurement of a large number of people.
Xovis Gender Statistics visualization in the Web UI
An AI-enabled sensor running firmware version 4.2.0 (and later versions) is required for the Gender Statistics plugin. The AI firmware runs on the following sensor models:
|AI-powered by default||AI-powered by free firmware upgrade|
|Sensor Type||PC2S / PC2R Standard||PC2S / PC2R - L||PC2S / PC2R - UL|
|Mounting Height Range [m]||2.4 – 4.0||2.5 – 4.0||2.4 – 3.25|
|Mounting Height Range [ft]||7.8 – 13.1||8.2 – 13.1||7.8 – 10.6|
- Activate the license on each sensor by adding the ID “PIGES” according to the AI Functions Quick Guide or the PC-Series User Manual
- Use the free 30-day evaluation to test the plugin
- Purchase an unlimited license code or use a recurring license through the Xovis Cloud
- Count lines: percentage of male and female forward and backward counts
- Object stream shows whether a track is from a female or male person
- WebUI visualization depending on privacy level
- Very low or high mounting heights (see compatibility chart) cause poor angles of view of faces and upper bodies
- Unfavorable lighting (too dark or too bright) limits accuracy
- Thermally challenging setups, such as direct sunlight on the sensor, direct mounting on isolating materials such as plaster ceilings etc., can lead to a throttled-down operation mode
- Potentionally reduced performance when combined with the Staff Exclusion plugin in very crowded settings
- Xovis SPIDER processing unit recommended for larger Multisensor setups (>9 devices) and for all settings more complex than a wide Multisensor counting line