Docstoc

An Application Of Morphological

Document Sample
An Application Of Morphological Powered By Docstoc
					   AN APPLICATION OF MORPHOLOGICAL
       IMAGE PROCESSING TO FORENSICS

ABSTRACT:-
                       Morphological Image      interest.        Morphological                   Image
Processing is an important tool in the          Processing       is used in the place of a
Digital Image processing, since that            Linear Image Processing, because it
science can rigorously quantify many            sometimes        distort           the       underlying
aspects of the geometrical structure of         geometric form of an image, but in
the way that agrees with the human              Morphological image Processing, the
intuition and perception.                       information of the image is not lost.
            Morphologic image processing                      In the Morphological Image
technology is based on geometry. It             Processing the original image can be
emphasizes        on    studying   geometry     reconstructed by using Dilation, Erosion,
structure    of    image.   We     can   find   Opening and Closing operations for a
relationship between each part of image.        finite no of times.
When        processing       image       with               The major objective of this
morphological theory. Accordingly we            paper is to reconstruct the class of such
can comprehend the structural character         finite      length        Morphological          Image
of image in the morphological approach          Processing           tool      in        a     suitable
an image is analyzed in terms of some           mathematical          structure          using    Java
predetermined geometric shape known             language.
as structuring element.                                       The Morphological                  Image
            Morphological processing is         Processing           is      implemented           and
capable of removing noise and clutter as        successfully tested in FORENSICS:
well as the ability to edit an image based      Fingerprint Enhancement and reduction
on the size and shape of the objects of         of   noise      in        finger     print     images.
                                                   2


                                                       applying these structuring elements to
                                                       the     data     using      different     algebraic
                                                       combinations,                one          performs
                                                       morphological transformations on the
                                                       data.
                                                                      The     Morphological          Image
                                                       Processing operations are applied for
                                                       binary images in
                                                              FORENSICS:                      Fingerprint
INTRODUCTION:-                                                  Enhancement and reduction of
                The Morphological image                         noise in finger print images.
processing is generally based on the
analysis of a two valued image in terms                DIGITAL IMAGE PROCESSING:-
of   certain       predetermined     geometric                           Digital image processing
shape known as structuring element. The                involves         the        manipulation        and
term morphology refers to the branch of                interpretation of digital images with the
biology that deals with the form and                   aid of a computer and it is an extremely
structure of animals and plants.                       broad subject and it often involves
            A very well suited approach                procedures,            which            can      be
for extracting significant features from               mathematically complex.
images      that     are    useful     in   the                   The central idea behind is quite
representation and description of region               simple. The digital image is fed in to the
shapes is morphological (shape-based)                  computer one pixel at a time. The
processing. Morphological processing                   computer is programmed to insert the
refers to certain operations where an                  data in to an equation or series of
object is Hit or Fit with structuring                  equations, and then store the results that
elements and thereby reduced to a more                 may      display       or    further     processed.
revealing      shape.      These     structuring       Digital image processing used to solve a
elements are shape primitives which are                variety of problems. Although often
developed to represent some aspect of                  unrelated, these problems commonly
the information         or the noise. By               require methods capable of enhancing
                                                  3


pictorial    information        for      human        Similarly successful applications        of
interpretation and analysis.                          image processing concepts can be found
                Employment of fingerprints            in astronomy, biology, nuclear medicine,
as evidence of crime has been one of the              law enforcement, and defense.
most important utilities in forensics,
since the date 19th century. Where there
                                                      IMAGE DATA BASICS:-
are no witness to a certain crime, finger                        An image refers to a      2-D
prints can be very useful in determining              light intensity function, based on these
the offenders.                                        2-D array of numbers the images are
            The impressions left on the               categorized in to three forms,
surface are called latent fingerprints, and
caused by the ridges on the skin. In most                 Binary Image,
cases, they are incomplete and degraded.                  Grey Tone Image and
The individual features that uniquely                     Color Image
identify a fingerprint are called minutiae.
Thus, the basic ridge pattern together                Binary Image:-
with the minutiae and their location on                          The image data of Binary
the finger print pattern uniquely identify            Image is Black and White. Each pixel is
a fingerprint.The Morphological Image                 either ‘0’ or ‘1’. A digital Image is
Processing will enhance the degraded                  called Binary Image if the grey levels
noisy and        /   or    incomplete    latent       range from 0 and 1.
fingerprints.                                         Ex: A Binary Image shown below is,
            Image         enhancement      and
restoration procedures          are used to                  1       1      1          1   0
process         degraded        images      of
unrecoverable objects or experimental                        1       1      1          1   0
results too expensive to duplicate. In
physics and related fields, computer                         0       1      1          1   0
techniques routinely enhance images of
experiments in areas such as high-energy                     0       1      1          1   0
plasmas     and      electron    microscopy.
                                              4


             (A 4* 5 Binary Image)                designed to identify the file as an image
                                                  with the specific format.

                                                  FITTING AND HITTING:-
READING THE TAG IMAGE
                                                            The    Structuring    Element        is
FILE FORMAT:-
                                                  positioned at all positions or possible
              Image processing involves
                                                  locations in the Binary Image and it is
processing or altering an existing image
                                                  compared        with    the    corresponding
in a desired manner. The first step is
                                                  neighborhood of pixels.
obtaining an image, while this may
                                                            The    morphological          operation
sound obvious, it is not a simple matter,
                                                  resembles a ‘Binary’ correction. Where
since usable image data is not readily
                                                  the operation is logical than arithmetic in
available.
                                                  nature.
              The programmer needs a
simple method of obtaining image data
                                                  Ex.: Suppose we have two 3 * 3
in a standard, usable format, called
                                                  structuring elements
image file format . The image file format
determine the image data storage and
also gives additional storage information
                                                                    1      1          1
with the pixel values.
                                                                    0      1          0
      The image file consists of a                S1 =              1      1          1
Header Segment and a Data-Segment.
The Header will contain, at the very
least, the width and the height of the                              1      1          1
image. Since it is impossible to display          S2 =              1      1          1
or process any image without knowledge                              0      1          0
of its dimensions.

         The     Headers   of   most   file       In a given image A, B, C are the three

formats begin with a signature or magic           positions       where   the    S1       and   S2

number. A short sequence of bytes                 Structuring       Elements      should        be
                                                  positioned.
                                               5


                                                   corresponding structuring element pixel
                                                   is ‘0’.)
                                                                 For the above example, S1
                                                   and S2 HIT the Image in neighborhood
                                                   ‘A’. The same holds true at ‘B’. But at
                                                   neighborhood ‘C’, only S1 HITS the
                                                   Image.
                                                              In this concept HITS corresponds
                                                   to Union and where as the FITS
           Binary Image used to test               corresponds to Intersection.
Fitting    and    Hitting   of   Structuring                  Further more it is possible to
Elements S1 and S2                                 replace the set operation Intersection and
                                                   Union by the Boolean operators ‘AND’
FIT:-                                              and ‘OR’.
          The structuring element is said to
FIT the image if, for each of its pixels           DILATION:-
that is set to ‘1’, The corresponding
image pixel is also ‘1’.
          For the above example, Both S1                            Dilation - grow image
and S2 fit the image at ‘A’ (Remember              regions
that structuring element pixels set to ‘0’
are ignored when testing for a fit). S2                         Dilation causes objects to dilate
fits the image at ‘B’ and neither S1 nor           or grow in size. The amount and the way
S2 fits at ‘C’.                                    that they grow depends upon the choice
                                                   of the structuring element [3]. Dilation
HIT:-                                              makes an object larger by adding pixels
          A structuring element is said to         around its edges.
HIT and Image if, for any of it pixels                           The Dilation of an Image ‘A’
that is set to ‘1’, The corresponding              by a structuring element ‘B’ is written
Image pixel is also ‘1’. (Here also we             as   AB. To compute the Dilation, we
ignore Image pixels for which the
                                                6


position ‘B’ such that its origin is at             For Binary Image:-
pixel co-ordinates (x , y) and apply the                     Dilation operation is defined as
rule.                                               follows,
                                                                     D (A , B) = A  B
                1 if ‘B’ hits ‘A’                   Where,
g(x , y) =                                          A is the image
                0 Otherwise                         B is the structuring element of the order
                                                    3 * 3.
             Repeat   for   all   pixel   co-                   Many structuring elements are
ordinates. Dilation creates new image               requested for Dilating the entire image.
showing all the location of a structuring
element origin at which that structuring            EROSION:-
element HITS the Input Image. In this it
adds a layer of pixel to an object, there
by enlarging it. Pixels are added to both                           Erosion - shrink image
the inner and outer boundaries of                   regions
regions, so Dilation will shrink the holes
enclosed by a single region and make the                        Erosion   causes   objects     to
gaps between different regions smaller.             shrink. The amount of the way that they
Dilation will also tend to fill in any small        shrink depend upon the choice of the
intrusions into a region’s boundaries.              structuring element. Erosion makes an
             The results of Dilation are            object smaller by removing or Eroding
influenced not just by the size of the              away the pixels on its edges [3].
structuring element but by its shape also.                      The Erosion of an image ‘A’ by
Dilation is a Morphological operation; it           a structuring element ‘B’ is denoted as
can be performed on both Binary and                 A Θ B. To compute the Erosion, we
Grey Tone Images. It helps in extracting            position ‘B’ such that its origin is at
the outer boundaries of the given                   image pixel co-ordinate (x , y) and apply
images.                                             the rule.
                                                7


                 1 if ‘B’ Fits ‘A’,                 For Binary Images:-
g(x , y) =                                                    Erosion operation is defined as
                  0 otherwise                       follows,
                        .                                     E (A, B) = A Θ B
                  Repeat for all x and y or         Where,
pixel co-ordinates. Erosion creates new             A is the image
image that marks all the locations of a             B is the structuring element of the order
Structuring elements origin at which that           3 * 3.
Structuring Element Fits the input                       Many structuring           elements    are
image. The Erosion operation seems to               required for eroding the entire image.
strip away a layer of pixels from an
object, shrinking it in the process. Pixels         OPENING:-
are eroded from both the inner and outer
boundaries of regions. So, Erosion will
enlarge the holes enclosed by a single                                   Opening    -    structured
region as well as making the gap                    removal of image region boundary
between different regions larger. Erosion           pixels
will also tend to eliminate small                              It   is    a   powerful    operator,
extrusions on a regions boundaries.                 obtained by combining Erosion and
             The result of erosion depends          Dilation.       “Opening       separates    the
on Structuring element size with larger             Objects”. As we know, Dilation expands
Structuring elements having a more                  an image and Erosion shrinks it [3].
pronounced effect & the result of                   Opening generally smoothes the contour
Erosion with a large Structuring element            of an image, breaks narrow Isthmuses
is similar to the result obtained by                and eliminates thin Protrusions [1].
iterated     Erosion   using    a     smaller            The Opening of an image ‘A’ by a
structuring element of the same shape.              structuring element ‘B’ is denoted as A
              Erosion is the Morphological          ○ B and is defined as an Erosion
operation, it can be performed on Binary            followed by a Dilation, and is
and Grey images. It helps in extracting             written as [3],
the inner boundaries of a given image.                       A ○ B = (A Θ B) B
                                              8


         Opening operation is obtained            Closing is obtained by doing Erosion on
by doing Dilation on Eroded Image. It is          Dilated image. Closing joins broken
to smoothen the curves of the image.              objects and fills in unwanted holes in
Opening spaces objects that are too close         objects.
together, detaches objects that are                    Closing involves one or more
touching and should not be, and enlarges          Dilations followed by one Erosion.
holes inside objects.
         Opening involves one or more             RESULT :-
Erosions followed by one Dilation.                FINGER PRINT ENHANCEMENT:-
                                                                      Fingerprints are unique.
CLOSING:-
                                                  The differences between fingerprints are
                                                  due to the type and the position of the
                                                  ridge characteristics. In most cases,
                Closing     -    structured       acquired latent fingerprints are degraded,
filling in of image region boundary               noisy and / or incomplete. Thus to
pixels                                            reduce the rejection rates during the
          It is a powerful operator,              matching stage, latent fingerprints have
obtained by combining Erosion and                 to be enhanced prior to matching. This
Dilation. “Closing, join the Objects” [3].        can be enhanced using Morphological
Closing also tends to smooth sections of          Image Processing.
contours but, as opposed to Opening, it                        The fig (a) is original image ,
generally fuses narrow breaks and long            to enhance the fingerprints we are
thin Gulf’s, eliminates small holes and           subjecting     to     the    Morphological
fills gaps in the contour [1].                    Operations. When the image is Dilated,
          The Closing of an image ‘A’ by          it leaves specific & clear ridges to
a structuring element ‘B’ is denoted as           visualize, can be seen in fig1.By Eroding
A● B and defined as a Dilation followed           the fig (a), the ridges are thickened for
by an Erosion; and is written as [3],             analysis. Can be seen in fig 2.
         A● B = (A  B) Θ B
                                              9


By performing Open operation to fig (a),          2. ERODED IMAGE:-
the ridges that are broken can be joined
to analyse the fingerprints clearly, can be
seen in fig 3. And by performing Close
operation to the fig (a), the ridges which
are overlapped can be separated and can
be analysed clearly, can be seen in fig 4.




a. ORIGINAL BINARY IMAGE :-
                                                  3. OPEN IMAGE:-




1. DILATED IMAGE :-                               4. CLOSE IMAGE:-
                                                10


CONCLUSION:-                                             Medical image analysis: Tumor

          This      report   represents   the               detection, measurement of size

practical operation of Morphological                        and shape of internal organs,

Image Processing and it successfully                        Regurgitation, etc.

performed     the      Fundamental        and            Robotics:       Recognition            and

Compound operations of Morphological                        interpretation of objects in a

Image processing on Binary images in,                       scene,     motion        control     and

    FORENSICS:                  Fingerprint                execution         through          visual

       Enhancement and reduction of                         feedback

       noise in finger print images.                     Radar imaging: Target detection
                                                            and identification.

IMPLEMENTATION:-                                     and this is further extended to Color
                                                     image concept and 24-bit True Color
                 This concept has been
                                                     concept and a special feature such as
implemented in java. The java platform
                                                     Automatic    selection     of     Structuring
provides a convenient representation for
                                                     element for object classification through
images that makes the implementation of
                                                     Morphology is still challenging to this
image processing software relatively
                                                     technique and have been chosen to be
straight forward.
                                                     the major direction of the future work.
        The Binary image operations
are implemented using Swings and have
a GUI for performing Dilation, Erosion,              REFERENCES:-
Opening & Closing operations                               www.encyclopedia.com
                                                           www.howstuffworks.com
FUTURE SCOPE:-                                             www.google.com
            The Morphological Image                        www.instrumentation.com
Processing can be further applied to a                     www.forensics.com
wide spectrum of problems including:                       www.imageprocessing.com

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:1/28/2013
language:Latin
pages:10