and why we use the term AR? And the things i found was
pretty interesting .First i thought that it is only used for gaming experience
. But i was wrong about the concept.
i found that it is used in various sectors.in the gaming
sector, medical ,mechanical, military etc. And it was a great experience for me
when i saw some people working with AR and VR. First i thought AR is some sort
of VR but i found it different.
In VR application the user is taken to a Virtual Reality and
he is not in the real world. He can’t work with VR in the real world. but in AR
the thing is different.AR adds more elements to the real world which does not
exist .The user can interact with both of the real word element and virtual element
.But the thing stack in my mind was how they identify the objects or the shapes
or the images ?i wanted to know is it possible to teach the AR application
about the basic shapes or it only works with the images of the objects
?Basically i’m interested about game development in AR .Because in recent time
people use too much time in playing mobile or pc games. Which they used to play
inside their houses .And it’s affecting their health. But when
“POKEMON:GO” was released it took out may people from their houses.It
also uses AR.I n this game it uses the user’s phone location and displays a
Computer generated animation on their phone scaren when they reached a certain
So thats why I choose the topic about AR recognizing 3D
Augmented reality is an addition to Virtual reality also
known as VR .In VR technology the user enters into another reality. Where the
user can’t interact with the real world, the user can’t see the real world
around him. But in AR the user doesn’t go to another reality.AR adds additional
virtual elements in the real world. the user can interact with both of the
virtual and the real elements.
There are many ways of recognizing a object in AR system.
Like image processing using 2D marker,3D marker, marker-less method and shape
recognition. But among above mentioned processes the shape recognition process
is advanced. Because in the image processing system the user can’t teach the
application about learning new objects. It can’t match objects beyond it’s
database’s predefined images. But in shape recognition process the user can
teach the software about new shapes by using some algorithms. And the main
facts is most of the objects shapes are same in world wide. And it can learn
about new shapes.
Augmented Reality is a live view of real world which
elements are augmented by a application like computer generated graphics ,image
,animation or sound. The field of AR exists for a long time and its growth is
also remarkable in recent time . In 1997,
the first author published a survey 1 (based on a 1995 SIGGRAPH course
lecture) in that paper the author wrote about the field, and he also
demonstrate the odds of AR and gathered the total summarization till that date
.After that paper had published there had been many development in AR sector.
Many seminars had been held after the 1990’s including the
International Workshop and Symposium on Augmented Reality 2.The basic goal of
an AR system is to enhance the user’s
perception of and interaction with the real world through supplementing the
real world with 3D virtual objects that appear to coexist in the same space as
the real world3.
There are many way’s to generate a AR system.One of the way
is to using 2D marker4.In this technic the object is identified by the visual
marker and using the video camera.the 2D matrix marker the technic uses is
squre shaped barcode that can identify 2^16 objects4.The application
binarized the captuerd.Then the binarized file is searched in the library and
Then the data is error checked.
Then there is another process in Image processing.The image
processing works in two ways.There is a process where the image of a phone is
matched in the database using image retrieval algorithm and also the GPS 5.In
this process the image is send to the database over broadband and GPS and by
using image processing algorithms the image is matched with the database.
In the other process the image is processed and matched with
the databse.where the databse is already imlemented in the mobile
phone.Research on image search services using mobile devices includes work by
Zhou et al. 6.There is also a set of vocabolary which is called
“visual-words”7.It helps to identify a single image in a large
database.A recent demonstration of an outdoor mobile augmented reality
application running on a cell phone is Nokia’s MARA project by Kähäri and
Murphy 8.in this demonstration no image processing is not done by the device
instate of that it matches the device’s location using the GPS and then show
the image of that location.kind of Google Map’s new image feature.
There is a another method called the marker-less image
processing techniques9.In the marker-less technic there are two phases one is
pre-processing and another real-time-processing.first, a 3D model of the
environment is reconstructed by the SfM method using photographs taken from
multiple viewpoints. The position and orientation data of all viewpoints are
stored in a database. Second, coordinates of 3D virtual objects that should be
augmented are defined relative to coordinates of 3DCG models of the environment
reconstructed by SfM and are stored in a database.Markerless AR depends on
the natural features of a surrounding rather than the fiducial identifying markers.
There are few problem with these above mentioned methodes.In
the 2D marker that it only identify 2^16 objects. despite of this the advantage
of this methode is that it is very simple and takes less memory4.
in the image processing and the it matches the image with
the data base.what if the image data base is not responding.the main problem
with this process is the angle of the image.if the image is taken in a angle it
may or may not be matched with the image of the data base. If the is not proper
broadband connection, then the image can not be sent to the data base and it
can’t receive data from the data base.
if we only use the GPS and there is a broad band error then the data
base can’t access the users location. the image process would be problematic if
the light is low. Beside it takes more memory than the 2D marker or marker
based motion capture. The marker-less process is bit much complex.
Another process is shape recognition.One of the best shape
recognition application is “Nestor”10.in Nestor when is video is feed it
takes each frame of the video which is passed through a filter to take shapes
which is tracked later.For the contour concavities each signature is then
calculated by the app for creating
hypothesis.Then each hypothesis is extracted from the library .The best thing
about “Nestor” is that the user can teach the application about shapes .It then
saves the shapes. A shape library is maintained by the system.which saves all
the shapes learned by the application.But the main problem is that it takes
more memory and processor for both mobile and computers.But the advantage is
that it can recognize the angular images.
The objective of the research topic is to discuss about the Spape
recognition process of the AR system and the differences between the other
processes of object recognition. And for this object I had to describe the old
used methods to object recognition and the shape recognition method . From this
research here raise some questions which are related to this research. This
question can give a new era of this sector. First of all How the AR works?
By analyzing the papers I found that AR
basically recognizes the objects either using shape recognition or the image
processing. Another one is why we should use the shape recognition rather than
the image processing? Basically in image processing the selected objects images
are already in the software’s database. And the objects doesn’t look the same
for which sometimes it would be difficult for the software to identify the
object but the shapes of the objects are same all over the world. So if all the
AR applications can identify the shapes differently it will be easy to identify
Proposed Research Methodology:
For gathering information and processing the research is
divided into some parts. One of the is which research type does the research
topic follow? Mainly the research topic follows the Descriptive and exploratory from Objective research.
The application of research study is divided into two types.
One is pure research and another applied research. In pure research theories
and hypothesizes are developed according to the topic and testing them. But in
my topic I have not developed any theory or hypothesis nor tested them. In
applied research the researcher try to solve some problems or try to answer
some questions. And I haven’t done it either. Because the the topic is new and
it is very difficult to do so.
In Inquire Mode the process try to find answer. It has two
– Structured approach
– Unstructured approach
The Structured approach is predefined .All the objectives,
design ,sample everything is predefined. The unstructured approach to inquiry
is usually classified as qualitative research.And my research is none of the
In objective correlational research the researchers tried to
find the interdependency between two or more aspects. Explanatory attempts to
answer why and how there is a relationship between two aspects? Descriptive
research describes problem, phenomenon, service systematically. Exploratory
research explores a area which is less discovered. As my research I tried to
describes Shape recognition process along with the other processes and try to
find the differences between the processes. And the AR area is less discovered
So my research follows the descriptive and exploratory research type.For this
research my goal is to make the shape recognition more useful and easy and I’ll
try to work on it.
For this research data should be primary data means the
researcher has to collect the data from the experiments. He can’t use others
data. In most of the research paper I found that every researcher is using
their own experiments data because each
of the mentioned method is different from each other. As I can’t undergo with
this experiments because I don’t have enough equipment and funding. But I went
through the other papers and saw their data and I found them amazing and their
1. R. Azuma, .A Survey of Augmented Reality,.
Presence: Teleoperators and Virtual Environments vol.
6, no. 4, Aug. 1997, pp. 355-385.
2. International Symposium on Augmented Reality.
Advances in Augmented Reality
HRL Laboratories, LLC
Naval Research Laboratory
Rockwell Science Center
Naval Research Laboratory
Georgia Tech GVU
4. Matrix: A Realtime Object Identi_cation and
for Augmented Reality
Science Laboratory Inc.
Higashi-gotanda, Shinagawa-ku, Tokyo 141 Japan