2008 10 15 Abstract n This paper investigates
2008 -10 -15 何寬宸
Abstract n This paper investigates the color facts from color images, obtains the relationship between chrominance and color components, and establishes a type of coordinate transformation which is able to improve chrominance of skin and lip. n With these coordinates, a new method of human face detection and location based on skin chrominance and lip chrominance transformation from color images is presented. n It is an effective and robust way to detect the position of objects which is not influenced by the pose of objects and their complex background. n The advantage of relatively stable chromatic information from color images is taken to detect or locate objects; also, the intensity information from color images is applied to enhance the contrast ratio between two objects with tiny differences.
Color space - YUV 1/2 n Expression to fit human visual property n n color space YUV through recoding the values of R, G, B. A color signal is composed of two parts n n n intensity and chrominance components. Y component represents the intensity of color U、V component two chromatic signals
Color space – YUV 2/2
Skin chrominance space and lip chrominance space n Skin color n n concentrated information in human's face color image more reliable than the features of other face organs detecting unclear images or small face images If the feature of skin color is stable n in other words if the invariance component of skin color can be extracted, then the changes from person to person or from bright to dark or different backgrounds can be eliminated. n The idea of detecting human objects in color images with skin color features can be accepted and possible. n Locating face with features from skin color has the advantage of rotation invariance and simplification.
Skin color and lip color clustered in RGB color space n Face skin colors have different appearances n n different races different persons or different individuals even the same person under different lighting source or wearing different clothes. However, they are in a relatively narrow color space. n n n investigated the distribution of face images under different video collecting devices different :sexes、 ages 、skin colors and found that human skin color and lip color are clustered in the RGB color space.
Relation of values of (R, G, B) on object
Eliminated intensity n The intensity can be eliminated from color space through normalization. n The pure color removed from the intensity is called chrominance, which can be represented by two vectors, U and V.
Hue phase distribution histogram of face skin color n 105~150
Rotation transformation
Skin、Lip chrominance coordinates n Lip: 85~102 (89)
Different results by applying variable Ф
Face location n To look for and locate human faces n we look for skin (including face, hand, neck and other things looking like skin) by applying the skin chrominance transformation n and then try to find lip or mouth (or lipcolor-like) by using the lip chrominance transformation
Restrained condition 1
Restrained condition 2 (1/4) n If each group satisfies restrained conditions, that region is probably the object region which we are looking for. n Next, smoothing the region of chromatic image r ( x, y)、g ( x, y)
Logarithm histogram
Restrained condition 2 (2/4)
Restrained condition 2 (3/4)
Restrained condition 2 (4/4)
Restrained condition 3 n In normal cases, the width and height of a human face accord with a given regular, that is a proportion in a range.
The result of detection and location by the above restrained conditions n Have done about 200 images from different channels with great difference in backgrounds
Perception peculiarity of intensity in eyes
Example of improving the contrast of skin and lip with suitable intensity coefficients
The results and conclusion (1/3) n All our experiments have achieved n n n We have detected n n rate of up to 26+ frames per second 24 bits color images. using Pentium-II 300 with CPE-1000 image board, JVC TK-1070 color camera MIMTRON MTV-3301 CB color camera 187 faces from 60 images with 196 persons under many complicated backgrounds by this approach, got to 95. 4% right detection rate by the approach. Also, it reveals good appearance in real time for any sequence of images with only one person.
The results and conclusion (2/3) n It is revealed that n human skin colors cluster in a small region in a color space; n each type of object has its own chrominance; n easily detecting one object can use its chrominance information by transforming from R, G, B to the chromatic coordinates in order to improve its chrominance; n it is effective and robust to detect a human face with skin chrominance and lip chrominance.
The results and conclusion (3/3) n We deduce two new coordinate types which can enhance skin chrominance and lip chrominance, respectively, and based on it we present the approach of human face detection and location. n Preprocessed by this enhanced transformation, the outline of interested objects is emerging obviously
Concluding summary n The advantage of this chromatic improving method is taking relatively stable information from color images to detect or locate objects: thus, it is not influenced by the pose of objects and illumination. n At the same time, the intensity information from color images is applied to enhance the contrast ratio between two objects with tiny differences.
- Slides: 28