Combining Haar Feature And Skin Color Based Classifiers For Face

Abstract

This paper presents a hybrid method for face detection in color images. The well known Haar feature-based face detector developed by Viola and Jones (VJ), that has been designed for gray-scale images is combined with a skin-color filter, which provides complementary information in color images. The image is first passed through a Haar-Feature based face detector, which is adjusted such that it is operating at a point on its ROC curve that has a low number of missed faces but a high number of false detections. Then, using the proposed skin color post-filtering method many of these false detections can be eliminated easily. We also use a color compensation algorithm to reduce the effects of lighting. Our experimental results on the Bao color face database show that the proposed method is superior to the original VJ algorithm and also to other skin color based pre-filtering methods in the literature in terms of precision. Combining Haar Feature And Skin Color Based Classifiers For Face.

HARDWARE & SOFTWARE REQUIREMENTS:
HARDWARE REQUIREMENTS:
  • SYSTEM :           Pentium IV 2.4 GHz
  • HARD DISK             :           40 GB
  • FLOPPY DRIVE :           44 MB
  • MONITOR :           15 VGA colour
  • MOUSE :          
  • RAM :           256 MB
SOFTWARE REQUIREMENTS:
  • Operating system :-          Windows XP Professional
  • Front End                   :-          Visual Studio .Net Frame .
  • Coding Language :-          C#
EXISTING SYSTEM:

Existing methods in the literature on face detection can be grouped as knowledge-based, feature-based, and template-based and appearance based methods. Face detection is an expensive search problem. In general, a sliding window is scanned through an image at various scales to classify the window as face or non-face. Therefore, many background windows need to be processed as well as actual face regions. The ratio of the number of non-face windows to face windows can be as high as 100000:1. Hence, a well trained classifier is necessary that will produce a low number of false positives.

PROPOSED SYSTEM:
  • We proposed skin color post-filtering method many of these false detections can be eliminated easily. We also use a color compensation algorithm to reduce the effects of lighting.
  • Our experimental results on the Bao color face database show that the proposed method is superior to the original VJ algorithm and also to other skin color based pre-filtering methods in the literature in terms of precision.
  • We propose a method that utilizes skin color detection to decrease the high false positive rate of the VJ face detector. The VJ algorithm uses only the brightness information in a search window, resulting in a high false acceptance rate due to face-like brightness patterns in the background. Therefore, skin-color is a complementary channel of information, and it is very fast to process.
  • We propose a skin-color based post-filtering method for color images. The windows that are detected as face by the VJ algorithm are verified if the window contains sufficient number of skin pixels.
  • Maximize the overall true detection rate, we adjust the parameters of the VJ algorithm such that the number of misses is low, and the number of false detections is high. Most of the false detections are easily eliminated by the proposed skin-color based post-filtering method.

Related Post