Face Detection in a Colour Image
Face Detection in a Colour Image
Muhammad Azeem Khan Raza
MSEE-17
Electrical Engineering Department
Military College of Signals, NUST
Rawalpindi, Pakistan. azeemraza1@yahoo.com Abstract— The perception of colour is very important for humans. Colour perception depends on both the physics of light and the processing by the eye brain which integrates the properties of stimulus with experience. Human use colour information to distinguish objects, food, places, materials, and even the time of the day
Keywords— Face detection, Color histograms, Color segmentation, Training samples, Circularity as variance of “radius”.
I. Introduction The presented MATLAB code detects the face in the colour image by locating regions of a colour image that is potentially a human face. First a training phase is undertaken to identify the nature of face pixels using samples from different people. It identify the pixel of the input image that are similar to the training samples and measures features such as area, elongation, circularity and then identify only those regions that are similar to human face.
II. face detection The color of a pixel contains good information for classifying that pixel in many applications. If a color model for human face color is described then it will help in finding the human face in a color image. Confusion is possible. For example pixel from brown cardboard can pass the skin color test and region shape might be needed to distinguish a polyhedral box face from an ellipsoidal human face. Two ways to detect the human face are.
• Color histogram • Color segmentation A histogram of a color image can be a useful representation of that image for the purpose of image retrieval and object recognition. A histogram counts the number of pixels of each kind and can be rapidly created by reading each image pixel just once and incrementing the appropriate bin of histogram. The intersection of
Muhammad Azeem Khan Raza
MSEE-17
Electrical Engineering Department
Military College of Signals, NUST
Rawalpindi, Pakistan. azeemraza1@yahoo.com Abstract— The perception of colour is very important for humans. Colour perception depends on both the physics of light and the processing by the eye brain which integrates the properties of stimulus with experience. Human use colour information to distinguish objects, food, places, materials, and even the time of the day
Keywords— Face detection, Color histograms, Color segmentation, Training samples, Circularity as variance of “radius”.
I. Introduction The presented MATLAB code detects the face in the colour image by locating regions of a colour image that is potentially a human face. First a training phase is undertaken to identify the nature of face pixels using samples from different people. It identify the pixel of the input image that are similar to the training samples and measures features such as area, elongation, circularity and then identify only those regions that are similar to human face.
II. face detection The color of a pixel contains good information for classifying that pixel in many applications. If a color model for human face color is described then it will help in finding the human face in a color image. Confusion is possible. For example pixel from brown cardboard can pass the skin color test and region shape might be needed to distinguish a polyhedral box face from an ellipsoidal human face. Two ways to detect the human face are.
• Color histogram • Color segmentation A histogram of a color image can be a useful representation of that image for the purpose of image retrieval and object recognition. A histogram counts the number of pixels of each kind and can be rapidly created by reading each image pixel just once and incrementing the appropriate bin of histogram. The intersection of
References: 1] Alasdair McAndrew, “An Introduction to Digital Image Processing with Matlab”, Victoria University of Technology: School of Computer Science and Mathematics 2] Rafael C. Gonzalez, “Digital Image Processing Using Matlab”, University of Tennessee. 3] Digital image processing by shapiro and stockman.