Muhammad been successfully applied on face recognition and

 

Muhammad Umer Khatab

Reg #. 1026-FBAS/MSCS/F17

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

International Islamic University Islamabad

Introduction

Object
recognition is an area of extensive research or a long period. The advance and
rapid development of computer hardware has enabled the usage of automatic
object recognition in more applications. These application includes industrial
image processing to medical applications. The object recognition is an umbrella
term of different algorithm designed for a large scale of applications, where
each application has its own some specific requirements and constraints. There
are many object recognition algorithm presented which concerned and categorized
into global approaches like, transformation search based methods, geometrical
modal driven methods, 3D object recognition methods, flexible contour fitting
algorithms and descriptor based methods 1.

3D
based recognition systems gained research attention recent years, especially in
the field of face recognition. It is the fundamental research in computer
vision. Mostly systems focused on facial characteristics data and used in
recognition purpose. The visual categorization and 3D object recognition is
very important in robot grasping with cameras. Incorporating 3D information can
potentially improve the performance of such a recognition system. 3D object
recognition is rapidly growing research area which based on used types of
features, 3D objects recognition methods can broadly divided into two main
categories, Global features based methods and local features based methods 1,
2, 4.

Kernel
methods are effective machine learning techniques for mostly image based
pattern recognition patterns. Incorporating 3D information is useful in such
types of applications. Such a kernel will naturally incorporate the 3D depth
information and can be use the system involving 3D object analysis and
classification. Kernel based methods have been successfully applied on face
recognition and cell classification. These methods construct the SVM kernels on
the wrapped phase data, which is preliminary raw data for many optical 3D
sensing methods, directly without explicit surface reconstruction and feature
extraction 4, 5.

Kernel
based methods play the very important role in the area of 3D object
recognition, because it varied application in automation, medical diagnosis,
surveillance and security systems, defense, content based image retrieval,
robotics and intelligence vehicle systems 2-3. Kernel based methods have some
benefits like, it avoids possible phase unwrapping errors introduced during
object reconstruction, and SVM can be very effective in solving classification
problems. Kernel methods can also use in regression, feature extraction, clustering
and other tasks. These methods construct the kernel directly from the optical
data without phase unwrapping. By avoiding the unwrapping, the resulting
kernels and algorithm will be more stable 4.

A distinct benefit of the kernel methods in machine learning is
the ability to harvest hidden patterns based solely on the natural similarity
measure which defined by the kernel, without using explicit feature extraction
4. Kernel methods work as when input image firstly segmented to get the
regions for interest and then from these the features are extracted and then
the dimensionality reduction methods are used and then the categorization
problems are solved by supervised classifier 2.

PICOC

Population

3D objects and face recognition.

Intervention

Kernel based method/technique, machine learning based
method, wrapped phase based method.

Outcome

Improve visual quality in resolution form.

 Context

3D objects and face recognition methods are use in
multiple application areas like, medical, robotics, security system and
intelligence vehicle system.

 

Q 1: Is there any literature available related to 3D object
recognition?

a.     What are the methods or techniques which based
on kernel for 3D object recognition?

Q 3: What measurement unit has been used for improvement of object
visual quality?

a.     What are the values obtain for improved
visibility of 3D object recognition using kernel based method?

b.     What is the priority of given method if other
methods or techniques are available?

Search
Strategy

Strategy for Deriving Search Terms

The
strategy used for deriving search terms is as below:

                     
i.       
Derive
major search terms from the research questions by identifying Population, Intervention,
Outcome, and Context:

Population

3D objects and face recognition.

Intervention

Method, technique, system, scheme, approach, way.

Outcome

Visual quality, resolution.

 Context

3D objects and face recognition methods are use in
multiple application areas like, medical, robotics, security system and
intelligence vehicle system.

 

                   
ii.       
Find
key words in the relevant papers. Note that the studies from which we have
taken the key words are not a subset of our primary studies. Rather they are
taken from the literature relevant to the disciplines 3-D object recognition.

Hong
Zhang and Hongjun Su 2009, Keywords-SVM

Phase
Uunwrapping; 3D Object Recongnition; Kernal Construction

Yulan
Guo 2014: Index terms

3D object recognition, keypoint detection, feature
description, range image, local feature

Ritu
Rani 2016: Keywords

Object recognition,
descriptors, moments, Neural Networks

Allah
Bux Sargano 2017: Keywords

Action recognition, deep
learning, transfer learning, hybrid classifier

 

                  
iii.       
Find
alternative spellings and synonyms for the search terms with the help of a
thesaurus. Also mention if a subject librarian and/or content experts in the
field are consulted.

3D objects

Triple entity, triplex items,3D things, three dimensional
objects  

Object recognition  

Object detection, object identification

Method

Approach, scheme, system, technique, way  

Kernel

Core, support vector machine (SVM)

           

                  
iv.       
Use Boolean OR to construct search strings from the search terms
identified in (i), (ii), and (iii).

3D
objects OR triple entity OR triplex item OR 3D things OR three dimensional
objects

Object
recognition OR object Detection OR Object Identification

Method
OR approach OR scheme OR system OR technique OR way

Kernel
OR Core OR support vector machine

 

                   
v.       
Use Boolean AND to concatenate the search terms and restrict the
research.

(3D objects OR triple
entity OR triplex item OR 3D things OR three dimensional objects) AND (Object recognition OR object Detection
OR Object Identification)
AND (Method
OR approach OR scheme OR system OR technique OR way) AND (Kernel OR Core OR support vector
machine)

 

Search
process and resources:

Search process for a systematic review should be rigorous and be
able to find as many relevant primary studies as possible. In order to make the
search process rigorous, this systematic review will consider two search
phases: primary/initial search phase and secondary search phase.

 

1.   
Primary Search Phase

The primary search phase will be directed
towards searching online databases, search engines, electronic journals,
conference proceedings, and grey literature. This section provides the details
of the electronic searches.

Online Databases:

·        
IEEE Explore

·        
Springer Link

            Online Search
engines:

·        
Google Scholar (scholar.google.com)

Primary search result

Serial No

Database Name

Number of Publication Found

1

IEEE Explore

32

2

Springer Link

3

3

Google Scholar

5

Inclusion and Exclusion Criteria

The study Inclusion based on
the given string which that contains the “3-D object recognition based on
kernel/SVM” keywords and the study that did not explicitly focus on kernel
based recognition methods and techniques were excluded.

Secondary Search Phase

The secondary search will be conducted to
complement the primary search phase. The following activities will be performed
in this phase:

1. The references for the articles
identified during the primary search phase will be reviewed. This process will
be iterative as the articles found relevant will be added to the list of
primary studies and this step will be performed for each of the identified
articles.

2. Citations will be reviewed for the
identified primary studies and the articles citing the identified primary
studies will also be reviewed by using backward and forward passes (as also
suggested by Webster & Watson 3). This process is also iterative in
nature.

3. Specific researchers will be
identified and contacted for advice on unpublished work and technical reports.

            Search Process
Documentation

Documenting the search process provides
transparency, helps prevent bias effect, provides details of thoroughness, and
enables replication of the search process.

 

 

Data Source

Documentation

Online
Databases/ Digital Libraries 

Name of
database
Search
strategy for database
Date of
search
Years
covered by search

Search
Engine

Name of search engine
Search strategy for the search engine
Date of search

Quality assessment checklist.

For
the assessment of study quality, checklists for qualitative and quantitative
studies have been developed. We have developed the checklists separately as we
expect to come up with both types of studies, which we believe would need
individual attention in terms of their quality assessment as the research
methodologies adapted in both types are different. Separate checklists also
help devise more detailed questions, which would not apply if using a single
common set of questions. These checklists are a means to assess the quality of
the selected studies and therefore their importance as evidence to answer the
research questions part of the systematic review.

Answer

Score

Yes

   
1

No

    
0

Partially

   
0.5

Quantitative checklist

The
quantitative checklist consists of 8 questions to be used for the evaluation of
quantitative studies and 3 questions to be used for the evaluation of
qualitative studies. The variables being measured adequately’ refers to whether
the measures/metrics were computed adequately on the collected data and ‘the
variables being validated adequately’ refers to the internal validity and that
the metrics are actually measuring what they are meant to measure.

Sr.
#

Question

Answer

1

The research question is clearly stated or
not?

Yes/ No/ Partially

2

Are the
kernel method clearly defined?

Yes/
No/ Partially

3

Is there any discussion about SVM (support
vector machine)?

Yes/ No/ Partially

4

Is the
described method/technique applied on 3D objects?

Yes/
No/ Partially

5

Is there any object detection method
described?

Yes/ No/ Partially

6

Is the
purpose of data analysis is clear?

Yes/
No/ Partially

7

Is the reporting clear and
coherent?

Yes/ No/ Partially

8

Are the data collection methods clearly described?

Yes/
No/ Partially

Qualitative checklist

Sr. No

Question

Answer

1

Does the study report clear, unambiguous
findings based on evidence and argument?

Yes/ No/ Partially

2

Are the data collection methods clearly described?

Yes/
No/ Partially

3

Are the links between the data,
interpretation, and conclusions clear?

Yes/ No/ Partially

Data Extraction Strategy:

After the primary studies have been selected and their quality
assessed, the data will be extracted. The data extraction forms and the
strategy to be adopted for recording the data are given in the sections below.

Data extraction forms are meant to contain all the information
that is necessary for answering the review questions and addressing the study
quality criteria.

Data Item

Value

Supplementary Note

Study
ID

 

 

Title

 

 

Author’s

 

 

Year
of Publication

 

 

Reference
Type

 

 

Publisher

 

 

Country
of Study

 

 

Application
Domain

 

 

Type
of Study

 

 

Article
Peer Reviewed

 

 

 

Data to extract and use to answer
research questions

 

Is there any literature
available related to 3D object recognition?

 

What are the methods or
techniques which based on kernel for 3D object recognition?

 

What measurement unit has
been used for improvement of object visual quality?

 

What are the values obtain
for improved visibility of 3D object recognition using kernel based method?

 

What is the priority of
given method if other methods or techniques are available?

 

 

Quantitative
Study Quality Assessment

 

The research question is clearly stated or
not?

 

Are the kernel method clearly defined?

 

Is there any discussion about SVM (support
vector machine)?

 

Is the described method/technique applied on 3D objects?

 

Is there any object detection method
described?

 

Is the purpose of data analysis is clear?

 

Is the reporting clear and coherent?

 

Are the data collection methods clearly described?

 

Does the study report clear, unambiguous
findings based on evidence and argument?

 

Are the links between the data, interpretation, and conclusions
clear?

 

Synthesis and Analysis of Extracted Data:

After the data have been extracted for
each primary study, they will be consolidated and summarized against each
research question. The information will be tabulated, as given below, in order
to make it more useful for further analysis and for finding the research gaps.

Question
1, 1(a)

            Question 1, 1(a) respectively
state:

                        “Is there any literature available related to 3D
object recognition?”

“What are the methods or techniques which based
on kernel for 3D object recognition?”

Study
ID

For
3D object recognition

Forecasting
technique/ method

Type
of technique (kernel based)

Notes
(if any)

 

 

 

 

 

 

 

 

 

 

 

 

 

Question 2, 2(a), 2(b)

            Question 2, 2(a), 2(b) respectively state:

“What measurement unit has been used for
improvement of object visual quality?”

“What are the values obtain for improved
visibility of 3D object recognition using kernel based method?”

“What is the priority of given method if other
methods or techniques are available?”

Study ID

3D object recognition

Forecasting technique

Type of technique (kernel based)

Measurement unit

Improved results

Technique priority

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

References

1

An Introduction to Object Recognition Marco Treiber
Siemens Electronics Assembly Systems GmbH & Co. KG Rupert-Mayer-Str. 44
81359 Munich Germany [email protected] [email protected]

2

A Comparative Study of Object Recognition Techniques
Ritu Rani HMR Institute of Technology and Management New Delhi, India
[email protected] Ravinder Kumar HMR Institute of Technology and
Management New Delhi, India [email protected] Amit Prakash Singh
University School of Information & Communication Technology Guru Gobind
Singh Indraprastha University Sector-16C, Dwarka, New Delhi, India

3

Human Action Recognition using Transfer Learning
with Deep Representations, Allah Bux Sargano, Xiaofeng Wang, Plamen Angelov
and Zulfiqar Habib  Department of
Computer Science, COMSATS Institute of Information Technology, Lahore 54000,
Pakistan

4

Wrapped
Phase Based SVM Method for 3D Object Recognition Hong Zhang and Hongjun Su
Department of Computer Science Armstrong Atlantic State University 11935
Abercorn Street, Savannah, GA 31419 USA

5

3D Object Recognition in Cluttered Sceneswith Local
Surface Features: A SurveyYulan Guo, Mohammed Bennamoun, Ferdous Sohel, Min
Lu, and Jianwei Wan

 

 

 

x

Hi!
I'm Rita!

Would you like to get a custom essay? How about receiving a customized one?

Check it out