Feature detection
An Important aspect of Image
Hasna O. & Mentor- Dr Man
A picture is worth more than a
thousand words
Image analysis (a.k.a. understanding),
image processing & computer vision plays
an important role in society today because
A picture gives a much clearer impression
of a situation or an object
Having an accurate visual perspective of
things has a high social, technical and
economic value
Digital image perception is used
Improving pictorial information for human
Processing of image data for storage,
transmission and representation for
autonomous machine perception.
Image Processing
Image Processing is to perform numerical
operations on higher dimensional signals such as
images and video sequences.
The objectives of image processing include:
– Improving the appearance of the visual data
image enhancement, image restoration
– Extracting useful information
image analysis, reconstruction from projection
– Representing the image in an alternate and possibly
more efficient form
transformation, image compression
Visible Human Project
The Visible Human
Project was an effort of
National Library of
Medicine (NLM) to build
a digital image library of
volumetric data
representing a complete,
normal adult male and
female. The data sets
released in 1994 and
Visible Human Project
What is Digital Image
The field of digital image processing refers to processing
digital images by means of a digital computer. A digital image
composes of a finite number of elements which have a
particular location and value and these elements are referred
to as picture elements, image elements, pels and pixels.
Digital Image: a two dimensional array of pixels.
–Size (or resolution) of an image:
width: N pixels, height: M pixels.
–Precision of pixels:
2n amplitude levels  n bits per pixel.
–Overall data file size: N  M  n bits
Image Example
LENA, 512x512, true color (24 b/p), 786432 bytes.
Images are referred to more than just the projections
generated by the visual band of the EM (electromagnetic) waves
apparent to humans. Images generated from the entire band of
the EM waves ranging from gamma to radio waves can be
perceived by imaging machines. Some of these images include
ultrasound, electron microscopy, and computer-generated
Image Examples
Size: 2048x2560
Graphic Example
Video Example
AKIYO, 352x288 (CIF), 24 b/p, 30 f/s, 72.99 Mbits/s.
Animation Example
There are 3 computerized processing levels:
Low-level process: - is characterized by the fact that both the
inputs and outputs are images. These involve primitive
operations such as image preprocessing to reduce noise,
contrast enhancement and image sharpening.
Mid-level process: - is characterized by the fact that its inputs
generally are images, but its outputs are attributes extracted
from those images such as, edges, contours, and the identity of
individual objects. Mid-level processing on images involves
tasks such as segmentation (partitioning an image into regions
or objects), description of those objects to reduce them to a
form suitable for computer processing, and classification
(recognition) of individual objects.
High-level process: - involves trying deduce an ensemble of
recognized objects, from image analysis to performing the
cognitive function usually associated with vision.
Origins of Digital Image
In the early 1920’s the Bartlane cable picture transmission
system was introduced, thereby reducing the transportation time
for a picture from New York to England by days. Digital images
were first applied in the newspaper industry when pictures were
first set by submarine cable between London and New York, then
the Bartlane cable was introduced reducing transmission time
from three weeks to three hours.
There were different phases in technology improvement; in 1921
the method used for receiving images through a coded tape by a
telegraph printer was abandoned in favor of a technique based
on photographic reproduction made from tapes perforated at the
telegraph receiving terminal with evident improvement in both
tonal quality and resolution.
The history of digital image processing is intimately tied to the development of
the digital computer. The first computers powerful enough to carry out
meaningful image processing tasks appeared in the early 1960s. This was
when there was significant development of the high-level programming
languages COBOL (common business-oriented language) and FORTRAN
(formula translator) and the development of operating systems. The birth of
digital image processing can be traced to the availability of advanced
computers and the onset of the space program during that period. Work on
using computer techniques for improving images from a space probe began at
the Jet Propulsion Laboratory in Pasadena, California in 1964 when pictures of
the moon transmitted by RANGER 7 were processed by a computer to correct
various types of image distortion inherent in the on-board television camera.
Fields that use Digital Image
In parallel with space applications, DIP is
used in
* Medical Imaging
* Earth Resources observations
* Astronomy
Medical Imaging (extracted from
Dr. Man’s Presentation)
Medical imaging is a rich field with multidiscipline nature,
involving nuclear physics, quantum mechanics, fluid
dynamics, advanced mathematics, biology, chemistry,
computer science and computer engineering.
It has become a primary component of modern
It is still a relatively new field with many unknown effects
and unanswered questions.
The technologies are evolving, and new equipment,
modality, study methodology have been constantly
Excellent opportunities for research and career
Medical Imaging
The invention in the early 1970s of computerized axial
tomography (CAT) is one of the most important events in the
application of image processing in medical diagnosis. Tomography
consists of algorithms that use the sensed data to construct an
image that represents a slice through the object which compose a
three-dimensional (3-D) version of the inside of the object.
X-ray CT
Tomography was invented independently by Sir Godfrey N.
Hounsfield and Professor Allan M. Cormack, who shared the
1979 Nobel Prize in Medicine for their invention. X-rays were
discovered in 1895 by Wilhelm Conrad Roentgen, for which he
received the 1901 Nobel Prize in Physics. These two inventions,
nearly 100 years apart led to some of the most active application
areas of image processing today.
Computer procedures are also used to enhance the contrast or
code the intensity levels into color for easier interpretation of Xrays and other images used in industry, medicine, and the
biological sciences.
Computed Tomography
Besides the natural images acquired from conventional
optical cameras, computer synthesized images become
more and more important in many application fields.
Non-invasive imaging modalities allow people to view
objects that can not be seen by human eye or camera,
– Internal organ of human body,
– Damaged part inside an airplane wing,
– Cloud covered city,
– Dark night battle field,
– Underground oil field…
Projection X-Ray
First X-ray
The hand of Mrs. Wilhelm
Roentgen: the first X-ray
image, 1895
For Mammography
Detection and diagnosis of breast cancer
based on some of the basis of these signs
made visible by Image Processing
Architectural distortions of normal tissue
Asymmetry between corresponding regions
of the images on the right and left breast.
Remote Earth Resources and
Geographers use the same or similar techniques to study pollution
patterns from aerial and satellite imagery. Image enhancement
and restoration procedures are used to process degraded images
of unrecoverable objects or experimental results too expensive to
duplicate. In archaeology, image processing methods have
successfully restored blurred pictures that were the only available
records of rare artifacts lost or damaged after being
In physics and other related fields, computer techniques routinely enhance
images of experiments in areas such as high-energy plasmas and electron
microscopy. Similarly successful applications of image processing concepts
can be found in astronomy, biology, nuclear medicine, law enforcement
defense, and industrial applications.
These examples illustrate processing results intended for human interpretation.
The second major area of application of digital image processing techniques
deals with machine perception. In this case interest focuses on procedures for
extracting from an image, information in a form suitable for computer
processing. Examples of the type of information used in machine perception are
statistical moments, Fourier transform coefficients, and multidimensional
distance measures. Typical problems in machine perception that routinely
utilize image processing techniques are automatic character recognition,
industrial machine vision for product assembly and inspection, military
recognizance, automatic processing of fingerprints, screening of X-rays and
blood samples, and machine processing of aerial and satellite imagery for
weather prediction and environmental assessment.
Fundamental Steps in DIP
Image acquisition- is the first process which involves
preprocessing such as scaling.
Image enhancement- this is bringing out obscured detail or
highlighting certain features of interest in an image. This
technique deals with a number of mathematical functions such as
the Fourier Transform.
Image restoration- it improves the appearance of an image but is
objective in the sense that this technique tends to be based on
mathematical or probabilistic models of image degradation.
Color image processing- this is used as a basis for extracting
features of interest in an image.
Wavelets- are the foundation for representing images in various degrees of
Compression- deals with techniques for reducing the storage required to save
an image, or the bandwidth required to transmit it.
Morphological processing- deals with tools for extracting image components
that are useful in the representation and description of shape.
Segmentation- partitions an image into its constituent parts or objects.
Representation and description- representation is necessary for transforming
raw data into a form suitable for subsequent computer processing. Description,
also known as feature selection, deals with extracting attributes that result in
some quantitative information of interest.
Recognition- assigns a label to an object based on its descriptors.
Feature Extraction- this is an area of image processing which involves using
algorithms to detect and isolate various desired portions of a digitized image or
video stream.
Image Enhancement
Histogram Example
Histogram Example (cont. )
Poor contrast
Histogram Example (cont. )
Poor contrast
Histogram Example (cont. )
Enhanced contrast
Smoothing and Sharpening
Smoothing and Sharpening
Image Analysis
Image analysis is to identify and extract useful
information from an image or a video scene, typically
with the ultimate goal of forming a decision.
Image analysis is the center piece of many
applications such as remote sensing, robotic vision
and medical imaging.
Image analysis generally involves basic operations:
– Pre-processing,
– Object representation,
– Feature detection,
– Classification and interpretation.
Image Segmentation
Image segmentation is an important pre-processing tool.
It produces a binary representation of the object with
features of interest such as shapes and edges.
Common operations include:
Thresholding: to segment an object from its
background through a simple pixel amplitude based
decision. Complicated thresholding methods may be
used when the background is not homogeneous.
Edge detection: to identify edges of an object
through a set of high-pass filtering. Directional filters
and adaptive filters are frequently used to achieve
reliable results.
Segmentation Examples
Edge detection
Feature Extraction
This is an area of image processing
that uses algorithms to detect and isolate
various desired portions of a digitized
What is a Feature?
A feature is a significant piece of
information extracted from an image which
provides more detailed understanding of
the image.
Examples of Feature Detections
Detecting of faces in an image filled with
people and other objects
Detecting of facial features such as eyes,
nose, mouth
Detecting of edges, so that a feature can
be extracted and compared with another
Feature Detection and
Feature Detection and Classification
• Feature detection is to identify the presence of a certain
type of feature or object in an image.
• Feature detection is usually achieved by studying the
statistic variations of certain regions and their backgrounds
to locate unusual activities.
• Once an interesting feature has been detected, the
representation of this feature will be used to compare with
all possible features known to the processor. A statistical
classifier will produce a feature type that has the closest
similarity (or maximum likelihood) to the testing feature.
• Data collection and analysis (or the training process)have
to be performed at the classifier before any classification.
Feature Extraction
Each line is represented by two parameters, commonly called r
and θ which represent the length and angle from the origin of a
normal to the line in question. In other words, a line is said to be
90° from θ and r units away from the origin at its closest point. By
calculating the value of r for every possible value of θ, a sinusoidal
curve is created which is unique to that point. This representation
of the two parameters is sometimes referred to as Hough space.
Thus the points to be transformed are likely to lie on an ‘edge’ in
the image. The transform itself is quantized into an arbitrary
number of bins, each representing an approximate definition of a
possible line. Each significant point (or feature) in the edge
detected is said to vote for a set of bins corresponding to the lines
that pass through it. By simply incrementing the value stored in
each bin for every feature lying on that line, an array is built up
which shows which lines fit most closely to the data in the image.
Hough Transform of Curves, and
Generalized Hough Transform
The transform described above applies to
finding straight lines; a circle for instance can be
transformed into a set of three parameters
representing its center and radius, so that the
Hough space becomes three dimensional.
Arbitrary ellipses, curves and shapes expressed
as a set of parameters can be found this way.
For more complicated shapes, the Generalized
Hough transform is used, which allows a feature
to vote for particular position, orientation
and/scaling or the shape using a predefined
look-up table.
Using Weighted Features
The Hough transform accounts for
uncertainty in the underlying detection of
edges by allowing features to vote with
varying weight.
Hierarchical Hough Transform
A final enhancement that is sometimes
effective is to perform a hierarchical set of
Hough transform on the same image, using
progressively smaller bins. If the image is first
analyzed using a small number of bins, each
representing a large range of potential lines, the
most likely of these can then be analyzed in
more detail. That is finding the bins with the
highest count in one stage can be used to
constrain the range of values searched in the
The data in this documentation has been quoted, cited
and extracted from the following sources:
Introduction to medical Imaging & Instrumentation –
Hong Man
The Art of Image Processing (operators and applications
– Hong Man
Digital Image Processing – Rafael C. Gonzalez, Richard
E. Woods
Handbook of Image and Video Processor
– Al Bovik
A New Approach to Image Feature
Detection with Applications – BS

Feature detection - Stevens Institute of Technology