ANNEXURE – I

University of Mumbai
FORMAT FOR SUBMISSION OF PROPOSAL FOR MINOR RESEARCH PROJECT

Part – A

1. Broad Subject: Information Technology

2. Area of Specialization : Automation System
3. Principal Investigator
(i) Name : Dr. Sunil B. Wankhade
(ii) Sex (M/F): Male
(iii) Date of Birth: 17/11/1968
(iv) Qualification: B.E., M.E. (Comp), Ph.D.
(v) Designation: Professor and Head I.T. Department
(vi) Address:

Office: Rajiv Gandhi Institute of Technology, Juhu Versova Link

Road Andheri (West), Mumbai-400053
Residence: C-101, Rainbow Apartment Sector-19, Airoli-400708
4. Co- Investigator (s)
(i) Name: Mr. Abhay Patil
(ii) Sex: Male
(iii) Date of Birth:08/04/1987
(iv) Qualification: M.E (Information Technology)
(v) Designation: Assistant Professor
(vi) Address:
Office: MCT’s Rajiv Gandhi Institute of Technology,
Versova, Andheri (W), Mumbai – 400053

Residence: - B-404, Pratik CHS, Sector 10, Airoli, Navi Mumbai -

400708
5. Name of the Institution where the project will be undertake:

(a) Department : Information Technology

(b) College / University : MCT’s Rajiv Gandhi Institute of Technology

6. Whether the College / University is approved under Section 2 (f) and 12 B of the
UGC Act?

Yes

7. Teaching and Research Experience of Principal Investigator

(a) Teaching Experience UG 21 Years

PG 02 Years

(b) Research experience

(c) Whether the project has been approved by the University for the Doctoral Degree? If
so, Please indicate
i. Date of Registration
ii. Name and designation of the supervisor (Guide) approved by the University.
iii. Name of the University where registered.
d) In case the teacher holds a Doctoral degree:
i. Title of the thesis: A Soft Computing Approach to Ant Based Routing in Mobile
Adhoc Network
ii. Year of the award of degree: 2015
iii. Name of the University: SGB Amravati University Amravati

(e) Publication

Papers Published
Accepted
Communicated

Books Published
Accepted
Communicated

(Please enclose the list of papers and books published and / or accepted during last
five years)
 Part-B

Proposed Research Work

8.
i. Project Title: Restoration of Image using only Noisy Data
ii. Introduction
 Origin of the research problem:
In this growing age of Image Processing, it is important for the
results to be crystal sharp and to be more informative. Nowadays,
images are taken by anyone, anywhere. Right from our mobile phones to
professional cameras, all are used to capture images. They have various
types of quality of hardware which affects the quality of the images.
These all images are susceptible to some kind of noise.
Image noise may be caused by different intrinsic (i.e., sensor)
and extrinsic (i.e. environmental) conditions which are often not possible
to avoid in practical situations. Therefore, image de-noising plays and
important role in a wide classification, where obtaining the original
image content is crucial for strong performance.
Digital images have played a vital role in research and
development of technology. One of the fundamental challenges in the
field of image processing and computer vision is image restoration,
where the underlying goal is to estimate the original image by
suppressing noise from a noise-contaminated version of the image.
The problem of image noise removal remains an open challenge,
especially in situations where the images are acquired under poor
conditions where the noise level is very high.
Image Restoration (IR) aiming to reconstruct a high quality
image from its low quality observation has many important applications,
such as low-level image processing. Medical imaging, remote sensing,
surveillance, etc.

 Interdisciplinary relevance:
Creating a neural network model which will have only corrupted
data as input and gives out a clean image which helps in improving
medical imaging and astronomical imaging.
This is a combination of two disciplines i.e. image processing
and Deep Learning. The image processing uses the concepts of Digital
 Signal Processing for reconstruction of image signal. These are straight
sections of Information Technology.

 Review of Research and Development in the Subject:

Mathematically, Image Restoration problem can be expressed as y =
x + n, where x and y are degraded image and the original image respectively,
and n denotes the additive noise. Restoring x from y is a challenging inverse
problem.
In the past a few decades the Image Restoration problems have been
extensively studied. However, they still remain as an active research area.
Generally existing Image Restoration methods can be classified into
main categories i.e. model based methods and learning based methods.
The model based methods attack this problem by solving an
optimization problem, which is often constructed from a Bayesian
perspective. In the Bayesian based models, the solution is obtained by
maximizing the posterior probability.
Different from the model-based methods that rely on a carefully
designed, the learning based Image Restoration methods learn mapping
functions to infer the missing high-frequency details or desirable high
quality images from the observed image. In the past decade, many learning-
based Image Restoration methods have been proposed. Inspired by the great
successes of the deep convolutional neural network (DCNN) for image
classification, the deep convolutional neural network models have also been
successfully applied to image IR tasks.
Learning based methods have shown better Image Restoration
performances than conventional optimization-based Image Restoration.

iii. Objectives
 To implement and develop a neural network model that would reduce
noise and recover clean image.
 To validate the results, the researchers performed experiments by
introducing different varieties of synthetic noise like:
i. Gaussian
ii. Poisson
iii. Bernoulli noise
 To train the network to remove text from the images

iv. Methodology and Plan of Work

Recent advances in the deep neural network models possible to restore
images by only looking at corrupted images, models can learn to map corrupted
observations to the unobserved clean images without explicitly training with the
clean images and without likelihood models of the corruption images.
In practice we show that a single model can learn photographic noise
removal, Gaussian Noise, Poisson noise, text removal - all corrupted by
different processes - based on noisy data.
 This can be done by training a regression based convolution Neural
Network Model with a large number of pairs (x, y) of corrupted images x and
clean target y, minimizing the empirical risk

Our base-line is a state-of-the-art method “RED 30”. A 30 layer

hierarchical residual network with 128 feature maps, which has been
demonstrated to be very effective in various image restoration tasks.
We train the network using 265*256 pixels crops drawn from the 50k
images in the IMAGENET validation set. For testing purpose, we use three
well-known datasets: BSD 300, SET14 and KODAK.

Plan of work
Planning the architecture of 24-8-2018  Which method to use for building a
the model: (1 month) to 24-9- neural network model.
2018  What person will do which work?
 Time allocation for various stages.
 Cost estimation.
 Resources required.

Data Preparation:(1 month) 25-9-2018  Download IMAGENET Database.

to 25-10-  Write a python script to create
2018 Gaussian, Poisson, and Bernoulli
based noisy data.

Training the model: (3 26-10-2018  Build a deep neural network model

month) to 26-1- using Tensorflow.
2019  Train model using Training data

Testing model on the Test 27-1-2019 Test model using testing data.
Data: (1 month) to 27-2-
2019

Deploying the model API: 28-2-2019 Deploy model as an API.

(15 days) to 13-03-
2019

i. Details of the Collaboration, if any intended: NONE

v. Conclusion:
 To show that simple general regression neural network can lead to new
capabilities in learning image recovery.
  It is possible to recover signals under complex corruptions without
observing clean images, without an explicit statistical characterization of
the noise or other corruption
 The classic BM3D algorithm that draws on self-similar patches within a
single noisy image, has significantly solve the above problem of image
de-noising. We show these high restoration performance using deep
neural network can likewise be achieved entirely without clean data all
based on some general purpose deep convolutional model.
 This points the way to significant benefits in many applications by
removing the need for potentially strenuous collection of clean data.

9. Financial Assistance Required (Budget): Rs. 45,000

- The computation involves adding noise intentionally to more than 50,000
images and training the model on those images.
Ideal system to get the results in minutes instead of days has following
specifications:
 8 NVIDIA Tesla P100 GPUs with 128 GB on VRAM
 40-core Intel Xeon E5-2686 v4 (Broadwell) CPU with 480+ GB RAM
This high end computer can be used via pay per use basis on Amazon Cloud
Service as EC2 P3 Instance. We will be requiring p3.16xlarge variant which will cost us
approximately - Rs 43,500
More information at: https://aws.amazon.com/ec2/instance-types/p3/

- Contingency and metro travel – Rs.1500/-

10. Whether the teacher has received support for your research project
from the UGC under Major, Minor, and the support for research of from
any agency? If so, please indicate
(i) Name of the agency from which assistance was approved
(ii)Sanction letter No and date under which the assistance was
approved
(iii)Amount approved and utilized
(iv)Title of the project for which assistance was approved
(v)In case the project was completed, whether the work on the
project has been published
(vi) If the candidate was working for the doctoral degree, whether
the thesis was submitted and accepted by the University for the
Award of degree

(A summary of the report /thesis in about 1,000 words may

please be attached with the application)
 (vii) If the project has not been completed, please state the
reasons

11. (a) Details of the project/scheme completed on the ongoing with the PI
(b) List of Facilities that will be extended to the investigator (s)
by implementing institution

Name of Year Total Equipment/Infrastructure

the started Completed Facilities obtained
Agency

(c) Available equipment & accessories to be utilized for the

project.

12. Any other information which the investigator may like to give in
support of this proposal which may be helpful in evaluating
 To certify that

(a) The University / College is approved under the Section 2(f) and
12- B of the UGC act and is fit to receive grants From the UGC
(b) General physical facilities such as furniture /space etc. are
available in the Department/ college
(c) We shall abide by the governing the scheme in case assistance is
provided to me/us from the UGC for the above project.
(d) We shall complete the project within the stipulated period. If I/we
fail to do and if the UGC is not satisfied with the progress of the
research project, the Commission may terminate the project
immediately and ask for the refund of the amount received by
me/us
(e) The above research project is not funded by any other agency

Name & Signature (a) Principal Investigator

Prof.A.E.Patil

(b) Co –Investigator

Prof. Govind P. Wakure

(c)Principal

Dr.U.V.Bhosle

(Signature with Seal)

 GUIDELINES FOR RESEARCH PROPOSAL FOR MINOR
RESEARCH PROJECT

Objectives:
To promote and encourage research culture among teachers so as to make them to
take up challenging research activities in future carrier.
Part A:

No specific Guidelines are required for this part (i.e. point 1 to 7)

Part B:

This part is related to the proposed research work. For this guidelines are as
follows:

8 (i) The proposal project should have brief title not exceeding one and half- lines to clearly
indicate its subject & aim.

8(ii) The introduction of the proposed research work must cover in the three topics given
as: Origin of Research Problem, Interdisciplinary Relevance, and Review of Research &
Development of the subject.

Provide background information on the project. Review the published work in the area with
appropriate references from national or international journals. Total number of references
should not exceed 25. Site the major references or reviews most pertinent to the subject and
justification to carry out the work in light of this background information.

The introduction should be written approximately of one page.

8 (iii) Objective: The rationale behind carrying out the project should be clearly written
as a justification for carrying out the research project. It should logically explain the
reason for carrying out the research work and explain how it will further add to our
knowledge in the respective field or result in further applications. This should be written
in half to three fourth

..9/-
 ..9..

page. Research work should not be repetitive of similar work done earlier either by the
investigator or any other person.

8 (iv) Methodology & Plan of Work: Briefly describe the proposed plan of work. Describe in
general the techniques to be used. Do not write details of the methodology in a stepwise
manner, Describe the facilities available in the laboratory to carry out the research work.

If facilities from some other laboratories are to be used, mention the type of facilities and
the name of the institution where these will be available.

Please confine your project so as a study & complete in-depth in a definite period i.e. 31st
March of the financial Year.

8 (v) Conclusions: The investigator should write a brief conclusion of the project which
should state how completion of the project will enhance our understanding of the concerned
subject and whether it will lead to any applications in the future. These should not exceed
half a page.

8(vi) Mention about collaborations, if intended.

9. Budget: The Budget should be prepared for different heads of expenditure such as

1. Consumables and Chemicals with information on their approximate costs.

2. Equipments: Request of small pieces of equipments may be considered.
Purchase of equipments of high prices through this project shall not be
considered.
3. Other expenditure such as contingency and local travel if needed may be
considered.

10 Write the details point wise.

11. Infrastructure facilities:

Workshop
Water & Electricity

..10/-
 ..10..

Laboratory space and furniture

AC room for equipment
Telecommunication
Transportation

Administrative and secretarial

support Library
Computation/ Documentation
Animal house/ glass house
Any other (please specify)
List the facilities that will be extended to the investigator (S) by the implementing institution.

11 (C) Available equipment and accessories to be utilized for the project.

Name of Equipment Model and make

Available with investigator 1.

2.

3.

Available in the Department 1.

2.

Available elsewhere in the 1.

Institution or in the region 2.
3.

..11/-
 ..11.

General Instructions:

The Minor Research proposal should be prepared & submitted strictly according to the
format given in ANNEXURE – I

The documents/ enclosed required with the proposals.

1. Photocopy of university approval from the CONCOL section obtained by

investigator in their colleges/ Institutions.
2. Endorsement (Annexure-II from the Head/ Principal / Director of the college /
institute on letter head)
3. If facilities from some other laboratories/ organizations are to be used, then a letter
of consent from the Head of that institution / organization allowing the use of such
facilities,
4. List of papers & books published /or accepted during last five years.

The Four copies of the research proposal (ANNEXURE – I) and soft copy in the form
of CD along with the above documents enclosed must be sent in one lot through proper
channel (i.e. through Director / Principal of the college / institute) to:

The Assistant Registrar,

Academic Planning & Development Unit,
Room No. 132, First Floor,
University of Mumbai,
Mumbai- 400 032.

Note that the proposal received up to 31th July 2011 will be evaluated & reviewed in
the month of August 2011. The proposals received after the 31th July 2011 will be
considered for the next Year i.e. 2012-13.

..12/-
 ..12..

While submitting proposal Principal Investigator should write following

information on COVER PAGE.

Sr Particulars Information
No.
1 Principal Investigator Name –
E- mail ID.-
Office Ph. No. –
Mobile No.-
2 Name of college
3 Program Arts/Science/Commerce/ Engineering/Law etc.
4 Title of the Project

Important Note:-

1) After completion of the research project, Principal Investigator should submit

the report of the same in two copies (hard copies) and one soft copy in the form
of CD to APD unit of the University of Mumbai within stipulated time.
2) Principal Investigator should give presentation of completed project before the
Expert Committee and Authority of University of Mumbai.
3) It will be highly appreciated if paper is written in National/ International
journal of high reputation.
4) Quality projects recommended by committee will be put on Website of
University of Mumbai
5) Best project Award of Rs. 5000/- (Rs. Five thousand only) will be given to the
project recommended by Expert Committee and Authority of University of
University of Mumbai.

..13/-
 ..13..

ANNEXURE – II

ENDORSEMENT:

Project title: --------------------------------------------------------------------------------------------

---
Certified that the institute welcomes participation of Shri./Smt. --------------------------------
------------------------------------------------------------------------- as the investigator for the
project.

Certified that the equipment and other basic facilities in the application and such other
administrative facilities as per terms and conditions of the grant will be extended to the
investigator (s) throughout the duration of the project.

Institute assumes to undertake the financial and other management responsibilities of

the project.

Institute assumes the responsibilities of submitting Grant Utilization Certificate Project

report and project completion report to the university not later that 30 th May of the next
financial year.

Name& Signature of the

Head of the Institution.

Place & Date:

Recommended

Principal of the College

..14/-
 ..14..

For favor of display on notice board

The Director
Board of College and
University Development,
University of Mumbai,
Mumbai-400 032.

CALL FOR GRANT OF RESEARCH PROPOSALS.

Objectives:

To promote and encourage research culture among teachers so as to make them to take
up challenging research activities in future carrier.

THE PROPOSALS SHOULD REACH THE UNIVERSITY UPTO 31st JULY 2011
ONLY, THROUGH / PRINCIPAL.

(For further details contact. -----------------------------------------------------------------------)

The name of Authority in the College / Institution.