Tag-based social image retrieval web interface

Abstract

This

project

is

a

development

of

a

web

interface

to

support

tag-­‐based

social

image

retrieval.

The

development

is

based

on

an

existing

system

named

CASIS,

which

was

written

is

java.

This

project

aims

to

provide

a

visualization

of

tag

relationships

as

well

as

the

image

ranking

through

a

web

interface

so

that

users

can

locate

the

target

image

more

effectively

and

easily.

The

main

challenge

of

this

project

is

to

understand

the

existing

system

and

eventually

reuse

the

original

components

as

much

as

possible,

for

the

purpose

of

consistency

and

accuracy.

To

enhance

the

user

experience,

a

different

approach,

tag

cloud,

is

chosen

to

visualize

the

tag

relationship.

The

development

also

used

some

external

resources

such

as

Twitter’s

Bootstrap

and

JQCloud

to

achieve

desired

results.

This

project

has

implemented,

in

the

form

of

a

JSP

web

application,

the

most

important

functions

in

the

original

system

including

formation

of

concepts

based

on

queries

entered

by

a

user,

and

the

images

grouping

based

on

each

of

these

concepts

as

well

as

a

way

to

customize

the

searching

methods

and

criteria.