Poised to realize the full potential of digital imaging, FotoNation’s FaceTracker is a face-detection and tracking technology now available in digital still cameras. The technology automatically detects the presence of a subject’s face and tracks the position of the face as it changes within the camera’s viewfinder. This information is then used to ensure the subject is in proper focus and properly illuminated for the final image capture.

The FaceTracker technology uses proprietary image processing algorithms to identify, lock, and track human faces as they move within the camera preview image frame. As a photo is being previewed in the viewfinder, the software detects and tracks up to ten faces simultaneously. Boxes are drawn around each face to provide feedback of correct detection (see the figure). Exposure settings are instantly applied, so that faces are never dark, even in poor lighting.

FaceTracker technology is currently available in Samsung and PENTAX digital cameras, and was recently announced for the Minton Brica DigiArt i90, Z870 and Z970 digital cameras. Additionally, according to FotoNation’s Eric Zarakov, the company plans to deploy the technology in digital cameras integrated into mobile phones.

Specifically, FotoNation now has a working demonstration application on the Nokia N series 73, N91 and N95, running on the host baseband processor. The performance is reported to be excellent, and the technology is not expected to introduce a significant power drain in these mobile devices.

Figure. FaceTracker technology can simultaneously identify and track multiple human faces—framed within superimposed lines—in the real-time image of a digital still camera viewfinder.

The software implementation of this technology used for these applications simplifies the hardware components. For example, the high-resolution available in the image sensors of digital still cameras is not a fundamental operating requirement for the technology. This is because only a very small number of pixels is needed to detect a face, according to Zarakov, so the technology is compatible with QVGA-level image sensors. Therefore, the technology is also highly compatible with mobile devices or other applications using these lower-resolution image sensors.

1 2 Next

RSS    Save to Del.icio.us  Digg This