This project presents a new approach for automatically tracking the human face as well as facial features (nose, mouth, eyes)in a clear way.
This technique became required in various future visual communication applications, such as teleconferencing, Facial recognition systems, Biometrics and Human computer interface etc.
The principle behind detecting the face feature is used to measure the respiration rate in the future as the nose represents the important region in the human face for breathing. Human face detection as elliptical area was investigated then image processing techniques were used to extract human face as elliptical area from the rest of image. Several techniques were applied to detect the nose inside the elliptical area as rectangle region and then the mouth and eyes regions were extracted inside the elliptical face area.
A skin-color segmentation with image processing techniques played an important role in detecting the human face as elliptic area and then several techniques were used such as enhancement, thresholding, Morphological, edge detections as well as binarization techniques to achieve the aims of the suggested methods.
Nose detection as a rectangle region was also investigated by looking for the longest vertical line in the elliptical area.
The nose was detected and extracted as rectangle region. Detecting the mouth was achieved by looking for the longest horizontal line under the tip of the nose then thresholding this region to detect the lips of the mouth; by extracting the points of the lips corners we extracted the mouth as elliptical region.
Finally, the eye regions were tracked in the upper part of ellipse above the tip of the nose and detected as rectangular regions.
Further work is in progress to enhance these techniques to take place in real time images as well as apply them in the medical field.