B. Android’s Software Development Kit (SDK)
SDK  is a set of software development tools which
aid in the development of Android applications. It includes
eclipse, SDK manager and Android Virtual Device (AVD)
manager. Eclipse is the Integrated Development Environment
(IDE) which includes important functions such as formatting
the code, graphical designing of the layout and different
perspectives such as Java perspective, DDMS perspective etc.,
which are very useful for coding and debugging. SDK manager
is a tool which is used to list out the installed Android
Application Programming Interface (API) and other tools and
provide a means to download and install other API’s and
MATLAB  is a FD processing unit which performs
image and video processing as it includes Image Acquisition
and Image Processing toolboxes. It also includes a Computer
Vision which has built in object descriptors and detectors.
These functionalities make it easier to implantation of an
algorithm for detection of various components in images.
V. PROPOSED TECHNIQUES IN FACE DETECTION UNIT
The Proposed techniques are aimed at reducing or eliminating
influence of external factors that hinder general FD system
A. Proposed HYCgCr Color Space
To detect facial region in the image from cluttered background,
this robust background removal technique is proposed.
This technique uses HSV  and YCgCr  color space.
HSV provides color information as Hue, Saturation and Value.
Hue refers to the depth of color of red, blue and yellow in
the range of 0 to 360. Saturation refers to the purity of the
color in the range of 0 to 100%. Value refers the brightness
of the color. YCgCr is derived from YCbCr  with a color
difference of using (G-Y) instead of (B-Y). YCgCr uses the
matrix expression as shown below.
color model in order to obtain the value of H, S, Cg and Cr and
then it fed to decision device to identify whether a particular
pixel is a skin or non skin region based on threshold values
fed to it.
B. Proposed Illumination Equalizer (IE)
The IE technique is a type of filtering that is used for
producing uniform illumination which would enhance FD
system. In the proposed method, pixel by pixel processing
is performed, some regions which are not skin region would
be identified as skin pixel due to noise. In order to detect
facial region, skin color pixel is assumed to occur in a cluster
of more than 1 or 2 pixels which can be used for reducing
the subject skin pixel. It is achieved by taking the difference
between the actual image and its opening of the image, which
is the collection of foreground parts of the image that fit a
particular structuring element.
The opening is performed by taking erosion followed by
dilation of the image. Erosion is carried out to remove the stray
pixels which enhance the boundary of the face region. The
noise pixels are moved and only cluster of skin color pixels
Fig. 2: Proposed FD process flow.
remains by the...