Image Processing by S.Steena Vaiz
May 11, 2015
Image Processing
by
S.Steena Vaiz
Introduction
Image processing is any form of signal processing for which the input is an image, such as photographs or frames of video; the output of image processing can be either an image or a set of characteristics or parameters related to the image.
Most image-processing techniques involve treating the image as a two-dimensional signal and applying standard signal-processing techniques to it.
Types
Image processing usually refers to digital image processing, but optical and analog image processing are also possible.
Image Processing Operations
Geometric transformations such as enlargement, reduction, and rotation
Color corrections such as brightness and contrast adjustments, quantization, or conversion to a different color space
Digital compositing or optical compositing (combination of two or more images). Used in filmmaking to make a "matte"
Interpolation, demosaicing, and recovery of a full image from a raw image format using a Bayer filter pattern
Image Processing Operations(Contd.) Image editing (e.g., to increase the quality of a digital
image)
Image differencing
Image registration (alignment of two or more images)
Image stabilization
Extending dynamic range by combining differently exposed images
Image Segmentation
Segmentation refers to the process of partitioning a digital image into multiple regions (sets of pixels). The goal of segmentation is to simplify and/or change the representation of an image into something that is more meaningful and easier to analyze.
Image segmentation is typically used to locate objects and boundaries (lines, curves, etc.) in images.
The result of image segmentation is a set of regions that collectively cover the entire image, or a set of contours extracted from the image.
Each of the pixels in a region are similar with respect to some characteristic or computed property, such as color, intensity, or texture. Adjacent regions are significantly different with respect to the same characteristic(s).
Image Processing Applications Computer vision Face recognition Feature detection Non-photorealistic rendering Medical image processing Microscope image processing Morphological image processing Remote sensing
Face Recognition
A facial recognition system is an image processing
application for automatically identifying or verifying a
person from a digital image or a video frame from a
video source. One of the ways to do this is by comparing
selected facial features from the image and a facial
database.
Face Recognition Used in :
Human and computer interface
Biometric identification
Principal Component Analysis (PCA) :
Widely adopted
Most promising face recognition algorithm
Solution : Applying PCA on wavelet subband
Subbands obtained using wavelet decomposition.
PCA applied on the generated subband face
Objective of Face Recognition : To determine the identity of a person from a given
face image. Complications occur due to variations in :
Illumination Pose facial expression Aging occlusions such as spectacles, hair, etc.
In the proposed method we proceed as follows :
Decompose face image into subbands using Discrete Wavelet
Transform (DWT)
Select mid-frequency subband (Diagonal) from third level.
Compute representational bases (apply PCA) for reference
images
Store as training image representations
Translate probe image into probe image representation using
representational bases
Use classifier to match with reference images to identify face
image
Discrete Wavelet Transform A face image of a person contains common (approximation) as well
as discriminatory (detail) information.
Discriminatory information is due to structural variations of the face.
The similarity information and discriminatory information are
segregated in different subbands at different levels of decomposition
of the face image.
Wavelet decomposition splits the facial features into :
Approximations, containing common (smooth) parts of the face
Details, containing the discriminatory (variations) information.
DWT (Contd.) The original image is decomposed into four
subbands - Approximation (A), Horizontal (H), Vertical (V ) and Diagonal (D) details.
where D = {H, V,D} such that A1= A2+D2 = A3+D3+D2.
Subband Creation & Selection
Principal Component Analysis (PCA) To recognize a face we need to measure the
difference between the new image and the original images
But the face contains an awful lot of data PCA is used to find a low dimensional
representation of data By means of PCA, one can transform each original
image of the training set into a corresponding eigenface
Eigenface
Eigenface is the eigenvector obtained from PCA Each eigenface represents only certain features of
the face In essence, eigenfaces are nothing but the
characteristic features of a face Similar faces (images) possess similar features
(eigenfaces) So, all images having similar eigenfaces are likely to
be similar faces
Face & their eigenfaces
Classification
An important part of image analysis is identifying
groups of pixels having similar spectral
characteristics and to determine the various features
This form of analysis is known as classification
Classification employs two phases of processing: Training – Create unique description based on
characteristic properties of image (face)
Testing – Match the description and classify the image
(face)
Face Recognition Process
The training & recognition processesTraining Process
Recognition Process
Training Stage Steps involved :
Apply 3-level Daubechies Wavelet Transform on
reference images
Choose subband 4 from level 3
Apply PCA on subband 4 & get eigenvectors and
eigenvalues
By arranging eigenvalues in a descending order,
eigenvectors with largest eigenvalues are used as
representational bases
Recognition Stage Steps Involved:
Apply 3-level Daubechies Wavelet Transform on
the test images
Apply PCA on subband 4 & get the eigenvectors
and eigenvalues
Use k-NN classifier to classify the test images into
appropriate classes based on the training set
Conclusion
Hence the nearest and similar neighbour is matched and the input face image is recognised using the Image Processing technique.
THANK YOU !!!