Face Detection Technology

Face detection technology was developed to automatically identify people's faces in a scene and make adjustments to achieve the optimum focus, exposure and white balance of a digital camera.

Face Detection Optimizes Image Quality

Face detection technology was developed to automatically identify people's faces in a scene and make adjustments to achieve the optimum focus, exposure and white balance of a digital camera. Examples of improving image quality by face detection are shown below.

In Figure 1 (a), because the camera is focused on the background, the background is clear but the doll in the foreground is out of focus and blurred. In Figure 2, exposure time and color are not properly adjusted for the people. The faces are overshadowed and the complexions are pale. When the face area is specified, focus can be adjusted as shown in Figure 1 (b), and when exposure time is adjusted properly, the complexion is optimized as shown in Figure 2 (b). This is why face detection is so advantageous in a digital camera.

(a) Face detection is not used   (a) Face detection is not used
(a) Face detection is not used   (b) Face detection is used
Figure 1:Face detection used for auto-focus

(a) Face detection is not used   (b) Face detection is used
(a) Face detection is not used   (b) Face detection is used
Figure 2:Face detection used for adjusting exposure and complexion

Capture the Moment

In a camera without the face detection function, the focus point is usually set at the center of the image. When people are not at the center, the focus point is set at the background and the people are out of focus. One method of solving this problem is for the user to first adjust the image frame so as to put the people at the center, keep pressing the shutter button in half-way to fix the focus point, adjust the camera to the desired angle, and finally fully press the shutter button. But this technique is awkward for many users.

Using face detection technology, an image with people's faces in focus is captured simply by pressing the shutter button, without any awkward technique requiring the user to press the shutter button half-way.

Practical Technology

Face detection with a digital camera involves two major issues: complicated scenes and processing time. The variety of possible scenes is virtually infinite. Moreover, objects may be blurred because they are out of focus. Face detection has to take into account a bewildering number of factors: the angle of a face, whether a face is partially hidden by a cap or glasses, facial expression, complexion, lighting conditions, background, and so on. A digital camera's processing capability is only 1/100 that of a PC, and the memory is 1/1000. Despite these constraints, face detection has to be executed within a reasonable time.

Page Top