1. Visual Commerce
Jay Shah
B.E Graduate of
Information Technology
D.J.Sanghvi College of
Engineering.
Tapan Desai
B.E Graduate of
Information Technology
D.J.Sanghvi College of
Engineering.
Pooja Shah
B.E Graduate of
Information Technology
D.J.Sanghvi College of
Engineering.
Prof.Vinaya
Sawant
Assistant Professor at
D.J Sanghvi College OF
Engineering
Prof.Anuja
Nagare
Assistant Professor at
D.J Sanghvi College OF
Engineering
ABSTRACT
Visual Commerce is a method of acquiring products
online via a search based on image. E-Commerce works
primarily on text based search. The primary purpose of
this paper is to suggest an alternative solution in the
form of Visual Commerce. Visual Commerce is an image
based search on the basis of the image recognized by
an application. Visual Commerce is a method which
uses image recognition and displays the results from
various online stores to the user.
The paper discusses the advantages of Visual
Commerce over the traditional E-Commerce. The paper
also explains the drawbacks of text-based search and its
limitations in the present internet scenario. The paper
further explains the future scope of visual commerce and
why it will be universal entry point in the consumer
process.
GENERAL TERMS
Image Processing[1], Visual Commerce, Visual Search,
E-commerce, M-commerce
KEYWORDS
Visual Commerce
1. INTRODUCTION
The digital world is getting crowded. The number of web
sites has skyrocketed in to the millions as companies
supplement their traditional merchandising avenues with
e-commerce. Only Americans spent $186 billion in
online transcactions while 85% of the Asian population
shop online. Retail success is no longer all about
physical stores. This is evident because of the increase
in retailers now offering online store interfaces for
consumers. There has been a boom in E-commerce
after the dot com burst. The net worth of Amazon.com a
top e-commerce website is 90 billion USD[2]. This clearly
shows how important online shopping is to a huge
population in the world.
Similarly, there has been a boom in visual search. Visual
search is a type of perceptual task
requiring attention that typically involves an active scan
of the visual environment for a particular object or
feature (the target) among other objects or features (the
distractors). Visual search could alter the search-
marketing playing field.
Visual Commerce is e-commerce based on visual
search. Visual commerce primarily targets mobile users.
Visual commerce could alter the search-marketing
playing field. Visual commerce will be critical for whoever
wants to get the attention of millions of mobile users,
who are projected to make up the majority of web traffic
within a few years.
2. DRAWBACKS OF TEXT BASED SEARCH
Text based search is the basic form of search since
many years. Humans are used to typing since the age of
type writers. So taking this fact into consideration text
based search has been the primary search technique in
all the e-commerce portals. But text based search come
with its set of limitations which are:
Quality of search: Everyone does not know what they
want to buy exactly. Some person may know the color,
while few may know the patter. However, they won’t
know the exact brand or category to search for in the e-
commerce portal. People scroll through thousands of
categories just to find the right match. Also typing the
patter or color in the text box hardly yields any results.
The entire process is time consuming and tedious.
Time constraint: Time constraint is one of the biggest
limitation in text based search. Individuals need to go to
the e-commerce portals and type exactly they want to
buy. This consumes a lot of time. It is lot easier just by
visual search.
Feasibility: Text based search is not feasible at all
times. It takes up the time and concentration of any
individual while typing. Also the results based on text
based search are not precise.

2. 3. RISE OF VISUAL COMMERCE WITH THE
RISE OF M-COMMERCE
The recent boom in mobile devices, in terms of both the
scope and nature of usage, is heightening the potential
for mobile commerce. A majority of world today have
access to a mobile phone; in fact, some have access to
more than one, and a growing number even rely
exclusively on wireless telephony. Recent models
integrate voice communications with various non-
voice functions, such as reading e-mail, viewing videos,
accessing the Internet, and using geo-location data. The
emergence and quick take-up of these models
underscores "the remarkable transformation the
cellphone has undergone since it was introduced over
25 years ago". At the same time as the transformation of
the wireless scene, the consumer payment market has
also been undergoing change. As illustrated by the U.S.
market the deployment and consumer uptake of new
payment technologies is accelerating. This is the context
in which businesses are creating strategies to engage
consumers in mobile commerce.
Visual commerce is a technology is useful for mobile
users who are on the go and seeking certain qualities in
a product. For example, someone with an iPhone could
see a shirt in a store and want to search for shirts with
similar quality over the net. He/she could use the phone
and search for it over the e-commerce website using
their mobile device. This eases up the process of typing
for the exact qualities of the product over the internet.
Instead the person gets an exact match of the patter he
desires by just using his device. Visual commerce is
quick and accurate.
4. THE PROPOSED WORKING
Visual commerce is basically visual search for e-
commerce. Visual search means searching for
something on the World Wide Web based on an input
image instead of text. Visual search currently is most
used in mobile devices. A visual search engine searches
images, patterns based on an algorithm which it could
recognize and gives relative information based on the
selective or apply pattern match technique.
Visual commerce works on the concept of mobile visual
search. A mobile image searcher is a type of search
engine designed exclusively for mobile phones, through
which you can find any information on Internet, through
an image made with the own mobile phone or using
certain words. Typically, this type of search engine uses
techniques of query by example or Image query by
example, which use the content, shape, texture and
color of the image to compare them in a database and
then deliver the approximate results from the query
using the e-commerce API’s.
The entire process is as follows:
First, the image is sent to the server application. Already
on the server, the image is then compared with the
existing data set in the database.
Once this whole procedure is done and the image is
recognized, the pattern or text found on the image is
matched in the e-commerce API in the form of query.
The results retrieved from the query are displayed in an
aggregated form to the user or the searcher.
Figure 1: Application detecting the object for visual
commerce
Figure 2: Application aggregating deals from all e-
commerce portals
5. ARCHITECTURE
Object Recognition Engine
An improved adaptive method of processing image data
in an artificial neural network has been developed to
enable automated, real-time recognition of previously
defined objects. The method involves a combination of
two prior object-recognition methods — one based on
adaptive detection of shape features and one based on
adaptive color segmentation — to enable recognition in
situations in which either prior method by itself may be
inadequate.

3. Figure 1: Object recognition engine architecture
This Optimal Adaptive Architecture involves interaction
between a shape-feature-based and a color-
segmentation-based method in a cyclic computation.
Using shape adaptive features and color adaptive
features from the previous cycle, an object and region of
interest containing the object are identified in the present
image by means of feature detection and color
segmentation. The region of interest is then used for
detecting objects from the sampling data.
Deals Aggregator
Based on the tags fetched from the user, this module
crawls the best deals from the top ecommerce portals
like Amazon, BestBuy, Walmart etc.
Figure 2: Deals aggregator architecture
6. EXISTING SYSTEMS
Google Goggles: Though Google Goggles is not a
dedicated visual commerce application the results
provided can help users buy products. The exact
working of google goggles is somewhat based on text
recognition. Whenever the user clicks a picture, the
application scan the images and retrieves the text. The
text is the sent as a query in the Google Search Engine
and the results are displayed. These results might
contain e-commerce portal links. Though these
application is not used for visual commerce it was one of
the very first advancement in the field of visual search.
The application only searches for the text from the
image, so it doesn’t work for random pattern or color
images which don’t have any text in them. Also the
results displayed are in the form of Google Search result
and the application does not redirect directly to the e-
commerce portal.
CamFind: It allows the user to take a photograph of any
object of interest and receive an identification of it and
simultaneously provides the user information about the
photographed item. CamFind is like Google Goggles, but
goes above and beyond. CamFind’s accuracy level is
above 85%, whereas Google Goggle’s accuracy is
between 15-20%. Moreover, CamFind is also capable of
identifying objects at any angle.
Once identification is returned to CamFind, the
application takes the user into the ‘intent screen’. Within
the ‘intent screen’ the user is presented with the
following options: price comparisons, similar images,
geo-location (closest place to find the item), as well as
internet search results.
CamFind also possesses the capability to scan and
identify barcodes, as well as QR codes. It also boasts a
translator and the ability to do a voice search. In
addition, a user can upload pictures from the iPhone’s
camera roll into the app for them to be identified.
CamFind’s best feature is that it reads for patterns or
color from any image clicked and gives items based on
those criteria’s. CamFind recognizes the entire object
instead of just the text.
7. ADVANTAGES OF VISUAL COMMERCE
Quick Search: The way visual search works is the
person doesn’t need to type or think of any keyword. The
person just clicks photos and gets results immediately.
Thus, Visual Commerce gives quick search results to the
user.
Reduce categorical search: With Visual Search the
picture which is clicked gives results. The user doesn’t
have to search categorically from thousands of articles.
Comparing prices online: The search engine gives the
best online deals available online for product in the
picture that is clicked.
8. DISADVANTAGES OF VISUAL COMMERCE
Sharp Angles: The photos to be clicked can only be
maximum at a certain angle. If the angle is too sharp the
results cannot be displayed. This delays the search.
Inaccurate detection of objects: Sometimes wrong
objects are detected. For example, if the person wants to
read lays it might detect something else.

4. 9. PROBLEMS IN IMPLEMENTING VISUAL
COMMERCE
Precision: Visual Commerce is still not precise. Visual
commerce can still be improved in many ways including
quicker recognition and better pattern matching.
Limited Dataset: The current dataset is limited for all
the phones. The servers can’t hold data related to
everything. These need large servers which increase the
cost of the application.
CONCLUSION
Visual Commerce is one of the latest advancements in
technology and will improve on Human Computer
Interaction. It’s a fine blend between Visual Search and
E-commerce, both of which are currently a booming tech
sectors. Though, Visual commerce isn’t that accurate, it
can be worked upon. Visual Search is still developing
and once the results are more precise, Visual
Commerce will be used every and anywhere!
REFERENCES
1. R. C. Gonzalez and R. E. Woods “Digital Image
Processing”, 2nd edition, Pearson Education, 2004
2. Visual Search and E-commerce:
orangecollarmedia.com