This was our interpretation on how the Google PLUS Imaging API works.Though name and content may not be close and descriptive enough, its purely an idea of how it might work.
Web & Social Media Analytics Previous Year Question Paper.pdf
Google enhanced imaging
1. Jaypee Institute of Information Technology,A-10 Sector 62 Noida
PHASE 2 : TOC PROJECT
Topic: Google+ Advanced Imaging Feature
Date : 23 September 2013
Submitted By : Udaysagar Kandpal
11103609
Siddharatha Dimania
11103587
Rajat Bothra Jain
11103541
Submited to : Mrs Shikha Jain
2. Introduction
Google Enhanced Imaging Feature : The Structure of Imaging Feature is divided into following parts
.
1. Dependent Modules
a. Browser(Client) and Google Server Module
b. Google Screen Interface
2. Main-Modules(Image Effects)
a. Auto-Enhance Module
b. Auto-Awesome Module
c. Auto-Highlight Module
Following pages elaborate on the dependent as well as main modules.
3. 1. Browser(Client) and Google Server Module
Process the request from
client
Send a request to Google Server via any Google Service
like
GMAIL/GOOGLE
ACCOUNT/GOOGLE
DRIVE/BLOGGER/YOUTUBE etc
If bandwidth available and connection is secure for
communication, server transmits and directs the login
page to display on the browser
Else
Signup data send to the
Google Server
If signup is selected
by the user
Display the Login page asking
for either signing up or signing
in
Else
Signup details valid
and correct
Fetch login
<db_name>
The Google database
searches for login and
matches the password
results
Sign in credentials submitted
by the user
from
<table_name>
Connection Error: The
content you are trying
to send is over an
unsecure connection
in
Server requests the database for
authentication
Result-Set not empty i.e. a valid credential was found
in the database
If the connection was found to be unsecure
during transfer of data
If the login was successful server redirects
Google Account Welcome Screen
for the Service (Google+ Screen
Interface Module)
4. 2. Google+ Screen Interface
else
If Upload photos is
selected
Upload Photos / Share
Photos / Create album
Browse for
files
If the file(s) is a
valid image
Photos is selected from
the links available in the
Google+ UI
Related
Modules
More than one photo
was selected
els
e
Do you want to
create
an
album?
Server processes
the file and adds
to its database
and checks the
file.
If one image file was
uploaded
Auto-enhance
Module
else
If enhance-image is
selected
If image contains a
human face
If all human faces are tagged
Add
(auto-enhance)
effects or done uploading
Add tags to the
faces
Edit image in creative kit /
Select done
5. 3. Auto Enhance
The auto-enhance feature has the following functionalities:
1. It will enhance and apply tweaks.
2. It makes simple adjustments like brightness, contrast, colour correction, noise reduction.
3. It also does the face-detection and automatically blur at the skin to hide zits (acne and pimples on
the faces).
4. All the adjustments can be made non-effective i.e. one can disable auto-enhance.
Server processes the
request
Upload a photo on
Google+ Photos tab
Server calls the Autoenhance related API
If Auto-enhance is defacto enabled and server
connection succeeded
If face is identified:
through algorithmic
analysis
If server connection did
not succeeded succeeded
Working
If server connection still not
succeeded but no timeout i.e.
after some time but before
timeout
If server connection did
succeeded and Autoenhance is de-facto
enabled
Identification of
pimples/acnes i.e. zits
If zits are present
on the faces.
Still Working
Remove the zits through
Image-Processing
Algorithms
else
Server Timed Out
Adjust the alpha value,
brightness (beta and
gamma correction)
Back to Google+
Screen Interface with
updated photo
6. 4.Auto Awesome
Features of Auto-Awesome are :
1 .Automatically combines photos together to create new ones.
2. It link multiple images together to create a motion gif, for instance.
3. It combines multiple images together where some people are smiling.
3. HDR processing is also an option, as is automatic panoramic images from ones taken side
by side.
4. All these features are all completely automatic and done for you without any intervention.
5. Google also goes through your photos and make a bunch of these for you automatically
over the past few weeks.
A few key about auto-awesome are given below :
HDR............
HDR, or High Dynamic Range, is the process of taking multiple exposures of the same image. By
merging these images together, your photos will achieve a greater range of shadows and light.
Uploading three similar images at different exposures--low, medium, and high exposure--will create
an HDR image for you through Auto Awesome.
Motion...........
If you've taken a series of photos in succession (at least 5), Auto Awesome will stitch these photos
together into a repeating short animation.
Smile.............
If you've taken a few group photos, Auto Awesome will choose the best shots of each person in your
image and merge them into one great looking photo.
Panorama..............
If you've taken a series of photos with overlapping landscape views, Auto Awesome will stitch these
photos together into a panoramic image.
Mix................
If you've taken a series of portraits sharing similar background elements, Auto Awesome will compile
these photos together into a photobooth style grid. Mix is meant to showcase portrait photos taken
with similar backgrounds in time, so it works best when there are close-ups of faces.
7. Switch the images to form a
panoramic image
Activate the auto
awesome feature and
process images
If count of photos
uploaded is more o
than or equal to five
If the images are
having overlapping
landscape views
If the images are having
overlapping landscape
views
If the images are shot in
quick succession of time
If the images
contain same
background
elements
If the images are having
overlapping landscape views
If photos are few and
has faces in it
Create a small gif animation
by successively combining
these photos
If the images
contain same
background
elements
If the images are shot in
quick succession of time
If photos are few and
has faces in it
Choose best shots of each
person in the photos and
create a large group photo
Combine the photo into
photo-booth style grid
If the image contain
same background
elements
If Process is
complete
If Process is
complete
If Process is
complete
Upload Photo to the Collections
If Process is
complete
If photos are
few and has
faces in it
8. 5.Auto-Highlight
Calls machine learning
algorithms + vector
support machine
Highlights tab
Duplicate photo is
discarded
If duplicates are present in the
photo
Select Albums Tab from Google+. At
Google+ Screen > Photos > Albums >
<album_name> > Share
if there are noises like red-eye in the photo or
borders are not smooth (detected by vector
support machines)
else
They are
prioritized in the
album
If smiling faces are recognized
in the photos/album
If the machine learning algorithms find important
land marks such as Eiffel Tower or friend that are
in your circles
The selected Photos are marked
as important
Pass through aesthetic value detector trained by
hundreds of human-raters
Unwanted/Not suitable photos
are removed before highlighting
Features of Auto-highlight that weren’t covered in the automata are :
Community-based Photo Ratings as Data Source
Visual Feature Extraction
Shape Convexity
Low Depth of Field Indicators
Region Composition
Size and Aspect Ratio
Red-eye is removed and non
smooth edges are blurred