Fig 4 - uploaded by Stephen Karungaru
Content may be subject to copyright.
Download
View publication
Detection results with the camera 1 meter from the faces. 

Detection results with the camera 1 meter from the faces. 

Source publication
Fig. 1. Skin color region (Face candidates) 
Fig. 2. Initial face locations detected inside the face candidates. 
Fig. 3. Adaptive search area control. Notice that the original face...
Fig. 4. Detection results with the camera 1 meter from the faces. 
Multiple Faces Detection in Real Time using Neural Networks
Article
Full-text available
  • Dec 2007
In this paper, a real time face detection method using several small size neural networks and a genetic algorithm with adaptive search area control is proposed. Neural networks and genetic algorithms may not be suitable for real time application because of their long processing times. However, in this paper, we show how fast speeds can be achieved...
Cite
Download full-text

Context in source publication

Context 1
... final system accuracy was calculated using the results from five minutes of testing. The number of frames processed was 8920 frames (Note that the total memory required to save these images is about 2GB and the total real time taken is about 20 minutes). About 30 frames are processed per second, five of which are not searched for faces. Therefore, the total frames to be searched for faces are about 7300. We carried out two sets of experiments, one with the camera at two meters (all five faces are visible), (System 1) and the other with the camera only one meter from the faces (only three faces within the camera view) (System 2), Fig. 4. The systems accuracy, false acceptance rate (FAR) and false rejection rates (FRR) are as shown in Table ...