Facial recognition is a hot topic. Most security cameras nowadays have enough processing power to analyze faces and compare them with those in a database. However, there are times when it is unnecessary or (for privacy reasons) undesirable to recognize someone using your security cameras. Instead, you might just want to know that a person is in view and, based on this detection, generate an event or alarm. This is known as face or person detection.

Facial recognition is useful when you want to track specific individuals or integrate it into youraccess control system. Facial recognition starts by detecting a face: an oval shape containing two eyes, a nose, and a mouth. The algorithm then compares thousands of measurement points on the face with the same points in the database. This step requires significant processing power.

If recognizing individuals is unnecessary or undesirable due to privacy concerns, you can configure the software in the security camerato detect faces and/or people. This allows you to quickly filter video footage, generate events or alarms, or limit the storage of footage by recording only when a face or person is detected.

What are the limitations?

Depending on why you want to use face detection, this technology also has its limitations. A security camera generates two-dimensional images, which means the camera can be fooled by photos or drawings of a face. If face detection is only used to trigger recording or to quickly search through footage, this isn’t a big issue; you might just get a few extra recordings. Another limitation is that covered faces cannot be recognized. Even headwear or scarves can cover the face enough to lower the detection rate, making it difficult to generate an event or alarm. Face detection alone is not suitable for alarming because there are simply too many “false positives” and “false negatives.” Ideally, you want only “true positives.”

Combined detection increases reliability

By combining face detection with person detection, reliability is significantly improved. Person detection is an algorithm that checks if the moving object is a human. It is a more complex algorithm than face detection but uses less processing power than facial recognition. Additionally, it doesn’t compromise privacy. The camera software can be set up so that an event is triggered when both a person and a face are detected. Person detection without face detection triggers a higher-level event because it suggests that someone may be hiding or covering their face, which could indicate suspicious behavior. Face detection without person detection sounds a bit eerie but usually points to a face used in an advertisement on a billboard or vehicle. No event or alarm is necessary in this case.

Conclusion

This combination of face and person detection can be a valuable addition to an existing security system. Many new cameras already have these algorithms built-in, making it easy to upgrade to next-level motion detection. Thanks to advancements in processing power and deep learning, these technologies are becoming increasingly reliable in the near future.