need help with networking project
PUT THE TITLE OF YOUR COURSE/SEMINAR HERE
by
Your Full Name
Submitted in partial fulfilment of the requirements for the degree of
MASTER OF ENGINEERING
Major Subject: Internetworking
at
© Copyright by your full name, year
Faculty of Engineering
Internetworking
The undersigned hereby certify that they have read and award a pass in INWK 6*** for the course/seminar entitled "This is the Title of Your Course/Seminar" by Your Full Name in partial fulfilment of the requirements for the degree of Master of Engineering.
___________________________
INTERNETWORKING PROGRAM
AUTHORITY TO DISTRIBUTE REPORT
Title:
Insert Title Here
The Internetworking Program may make available or authorise others to make available individual photo/microfilm or soft copies of this report without restrictions after insert date here.
The author attests that permission has been obtained for the use of any copyrighted material appearing in this report (other than brief excerpts requiring only proper acknowledgement in scholarly writing) and that all such use is clearly acknowledged.
Full Name of Author:
Signature of Author: _________________________
Date: _________________________
Table of Contents
LIST OF TABLES vi LIST OF FIGURES vii LIST OF SYMBOLS AND ABBREVIATIONS viii ACKNOWLEDGEMENTS ix EXECUTIVE SUMMARY x 1 INTRODUCTION 1 1.1 Set Up The Word Processor For Technical Writing 2 1.2 Acceptable Reports 4 1.3 Now For Your Report 4 1.4 Meaningful Methodology Title Goes Here 4 1.5 If This Is A Design Seminar Put A Meaningful Sub-Title 4 1.6 Put In a Section on the Results 5 1.7 Outline 5 2 PUT A MEANINGFUL TITLE HERE RELATED TO THE SPECIFICS OF THE SEMINAR 6 2.1 Top Of Hierarchy In Principles Of The Topic 6 2.2 Next Level In Hierarchy To Dig Deeper Into The Topic 7 2.2.1 More Detail On A Topic At This Level In The Problem 7 2.2.2 Put In As Many Sub-Sections As Required 7 2.2.2.1 In Fact To As Many Levels As You Wish 7 2.3 A Variation Of Relevance In The Problem 7 2.3.1 Another Point Of Interest To The Understanding Of The Problem 7 2.3.1.1 The Equations 8 2.4 Another Variation Or Points Of Interest To The Seminar 8 2.4.1 Filter Banks 8 2.4.2 Linear Prediction Coding 9 2.5 Summary 9 3 METHODOLOGY 10 3.1 Subsection Three-one 10 3.1.1 Title For Sub-Section Three-one-one 10 3.1.2 Title 10 3.2 Major Section Title 11 3.2.1 Sub-title With Heading Format 3 11 3.3 Summary 11 4 PUT A TITLE RELATED TO DESIGN OR ANALYSIS 12 4.1 The Perceptron 12 4.2 Neural Network Architectures 13 4.2.1 RBF Neural Network 13 4.2.2 RBF Training Algorithms 13 4.3 Summary 13 5 PUT A TITLE HERE RELATED TO IMPLEMENTATION 14 5.1 Introduction 14 5.2 Title 15 5.2.1 This stage 15 5.2.2 Sub Section Title 16 5.3 Summary 16 6 PUT IN A TITLE RELATED TO EXPERIMENTS AND RESULTS 17 6.1 Experiment One 17 6.2 Experiment Two 17 6.2.1 Evaluation Set 17 6.2.2 Recognition Set #1 18 6.2.3 Recognition Set #2 18 6.2.4 Boundaries Detection 18 6.3 Results Of Recognition Set #1 18 6.4 Summary 19 7 CONCLUSIONS AND RECOMMENDATIONS 20 7.1 Conclusions 20 7.2 Recommendations 20 REFERENCES 21 BIBLIOGRAPHY 23
LIST OF TABLES
Table 5.1 Mel-scale Frequency Bands and the Corresponding FFT Coefficients 14 Table 5.2 Phonetic Decomposition of Digits [Rab1993, Table 2.3] 15 Table 6.1 List of Speakers 16 Table 6.2 Average Recognition Rates in Manual and Automatic Systems 18
LIST OF FIGURES
Figure 2.1 A Speech Model [Rab1993] 5 Figure 2.2 Frames and Windows 7 Figure 2.3 A Typical Filter Bank 7 Figure 2.4 FFT and LPC Spectra 8 Figure 4.1 a) A perceptron, b) A multi-layer neural network 11 Figure 4.2 A Radial Basis Function Neural Network 12 Figure 5.1 A Block Diagram of the Proposed System 13 Figure 5.2 Endpoints of the Digit “five”. 14 Figure 6.1 Examples of Extra Subwords 17
LIST OF SYMBOLS AND ABBREVIATIONS
ABD Automatic Boundaries Detection
ANN artificial neural network
CD-PLU context-dependent phone-like units
CD-ROM compact disc
CI-PLU context-independent phone-like units
DTW Dynamic Time Warping
FFT Fast Fourier Transform
HMM Hidden Markov Model
Hz hertz
LMS Least Mean Squares algorithm
LPC Linear Prediction Coding
OLS Orthogonal Least Squares algorithm
PLP perceptual-based linear prediction
PLU phone-like units
RBF Radial Basis Function
SVF Spectral Variation Function
TIDIGITS Texas Instruments digits database
VQ Vector Quantization
ZC zero-crossing
ACKNOWLEDGEMENTS
I would like to express my
EXECUTIVE SUMMARY
The executive summary provides the rationale for the project/course/seminar together with the seminar's objectives and outcomes.
It also provides a brief overview of the methodology used to meet the objectives.
Finally, it must present a brief report on the outcomes in terms of the objectives.
The executive summary should be one page long and, like the references/bibliography, is to be single-spaced.
1 INTRODUCTION
The first page of chapter one is page 1, however it has no page number printed on it. All of the following pages must have the page number in the Header, top right. Use the View - Header and Footer to ensure the correct format and numbering.
Paragraph text is left justified with line spacing 1.5 and 12 point spacing before the first line i.e. p-text formatted which you see now in the top left of the screen. By using p-text formatting you will get the correct paragraph spacing and correct spacing between major section titles formatted as Heading 2. The spacing is to be one p-text formatted line feed.
The following is a typical organisation of a report and should be adjusted according to the topic and the number of chapters. You should be using main.dot as the template file.
The sections are then 1.1, 1.2 and so on and cover the following:
· Reasons for the project and an explanation of why the seminar is important
· the specifications and objectives of the seminar
· the achievements of the seminar with respect to the stated specifications and objectives
The last section must be an outline of the rest of the report on a chapter by chapter basis.
Put in reference citations [Alg1993] in the correct format in the body. Insert them alphabetically in the REFERENCES according to whether it is a text book [Rab1993], a journal article [Pic1993], or a Web Site reference (with date of last access) [Cis2000]. Examples may be seen in the REFERENCES section.
1.1 Set Up The Word Processor For Technical Writing
[Note the headings (except Heading 1) should all be done in "Title case" under "Format - Change Case".
Titles may only use abbreviations if the abbreviation has been assigned earlier in the
On average the report should have a Flesch reading ease value of a least 65%. The Flesch-Kincaid Grade Level of the report should be at least grade 8. Note that these two numbers tend to move in opposite directions. Therefore, a technical passive writing style can result in a reading ease value below 60% for grade levels of 7 or higher. For purely technical portions of the report, the grade level will probably be high and the reading ease value low. Check this paragraph with the Microsoft tools to see if you have set up your word processor for Canadian English. Also, check the Flesch ratings to ensure if you have enabled the ratings.
To check highlight the previous paragraph, use Tools, Spelling and Grammar, and say no when asked whether or no to continue checking the rest of the document.
1.2 Acceptable Reports
Here are eample of an unacceptable sentences. Word will identify errors automaticaly as you type by the use of wavy red or green underlines. And there is no checking that you have put the 2 required spaces between sentences. The previous sentence actually has an error as you are not permitted to begin sentences with the words "And, But, Hopefully".
1.3 Now For Your Report
Your first section of Chapter 1 should explain the background to the problem. In this section the specifics of the current seminar can be outlined. These specifics can comprise some combination of the following depending on the actual seminar:
· the detailed specifications to be met
· the current network under analysis
Again, references must be used in the correct format as in the reference to [Tos1996].
If there are a number of major sections of work then use sub-sections to describe them.
1.4 Meaningful Methodology Title Goes Here
A brief description is required of how the work was done to collect the appropriate information (or whatever) to meet the objectives.
1.5 Put A Meaningful Sub-Title
If this is a design seminar then any important considerations that affect the design should be briefly discussed in a sub-section of Chapter 1. The points made here will come up again in a later chapter devoted to a more detailed analysis and comparison of designs.
1.6 Put In a Section on the Results
1.7 Outline
[Note: The final section of Chapter 1 must be the report outline]
The material in this report is organised into # chapters and # appendices. The chapters deal with the following topics:
Chapter 2.....
Chapter 3.....
Chapter 4.....
Chapter 5.....
Chapter 6.....
Chapter 7 discusses the results of all experiments and draws conclusions. The chapter also provides some recommendations and suggestions for any future work.
Appendix A.....
2 PUT A MEANINGFUL TITLE HERE RELATED TO THE SPECIFICS OF THE COURSE/SEMINAR
Before the first heading x.1, each chapter from here on must start with a brief (one or two paragraph) statement of what is in the chapter.
2.1 Top Of Hierarchy In Principles Of The Topic
A discussion of the underlying principles is required. It is important to use references in the correct format as shown here [Mar1995][Rab1993][Rab1978].
Use bulleted lists where appropriate:
· Fricative (unvoiced) sounds: are produced by forming a constriction somewhere in the vocal tract and forcing air through the constriction so that turbulence is created resulting a noise-like excitation.
· Use automatic figure numbering and always refer to your figures in the text body. See Figure 2.1.
Figure 2.1 A Speech Model [Rab1993]
2.2 Next Level In Hierarchy To Dig Deeper Into The Topic
Regardless of the seminar the report must lay a solid foundation and introduce the reader to the problem's background in ever increasing levels of detail.
2.2.1 More Detail On A Topic At This Level In The Problem
Use sub-sections to present the problem details as necessary. Sub-section titles are to be indented by increasing amounts as the depth increases.
2.2.2 Put In As Many Sub-Sections As Required
The structure of the report must lead the reader through the detail in a hierarchical manner.
2.2.2.1 In Fact To As Many Levels As You Wish
This should be deep enough for most people
2.3 A Variation Of Relevance In The Problem
It is important to reveal all the details of the problem to the reader. This chapter is used to make the reader aware of all the nuances of the problem so that any subsequent trade-offs or measurements are justifiable in the context of the problem description and objectives.
2.3.1 Another Point Of Interest To The Understanding Of The Problem
Remember to embed figures and to refer to them in the text as shown in Figure 2.2. Give references as required [Pic1993].
2.3.1.1 The Equations
(2.1)
(2.2)
(2.3)
2.4 Another Variation Or Points Of Interest
The output spectrum of...
2.4.1 Filter Banks
Another popular analysis method of speech is the filter bank model [Pic1993][Rab1993]. A typical filter bank structure is illustrated in Figure 2.3.
Figure 2.3 A Typical Filter Bank
2.4.2 Linear Prediction Coding
The linear prediction coding (LPC) performs spectral analysis on frames of speech. Figure 2.4 shows FFT and LPC spectra of a speech signal.
Figure 2.4 FFT and LPC Spectra
2.5 Summary
Each chapter must conclude with a summary to recap the major points of the chapter. Also, make sure that the pages run correctly numbered with no blank pages as the use of the main document sometimes causes an extra page to be inserted.
3 METHODOLOGY
Place a brief statement of what this chapter covers and how it follows from the preceding chapter here. Remember to use references as appropriate in this format [Gup1991][Sve1989]. Remember that a paragraph consists of a number of sentences.
Remember that when you refer to a chapter with its number you use a capital as in Chapter 3.
Paragraphs are to be separated by 12 points so use p-text for paragraphs. Use of personal pronouns is not permitted – use an impersonal writing style throughout the report. Use left paragraph justification throughout the report.
One p-text formatted empty line must be inserted before major section headings.
3.1 Subsection Three-one
Remember to structure your report into Sections and sub-sections. The topic of the section has to be logically divided into distinct sub topics so that you can do the structure correctly and meaningfully.
3.1.1 Title For Sub-Section Three-one-one
Sub-sections need not be very long, however they should consist of a number of paragraphs rather than a single paragraph (otherwise sub-sections are the incorrect way to present your information.
3.1.2 Title
Use as many sub-titles as are necessary.
3.2 Major Section Title
A few relevant paragraphs related to the topic of the title should be used.
3.2.1 Sub-title With Heading Format 3
Remember to use as many sub-titles as is necessary.
3.3 Summary
A summary starts here.
4 PUT A TITLE RELATED TO DESIGN OR ANALYSIS
The report requires design and analysis of the specific problem and this must be presented. All trade-offs must be discussed as well as performance issues. The underlying technical principles can be laid out in the major sections.
4.1 The Perceptron
Figure 4.1 a) A perceptron, b) A multi-layer neural network
The function of the perceptron can be formulated as:
(4.1)
or equivalently
(4.2)
4.2 Neural Network Architectures
Neural networks used today can be classified in
to many over-lapped classes according to [Hus1993][Lip1987]:
4.2.1 RBF Neural Network
A Radial Basis Function (RBF) neural network is a two-layer network as shown in Figure 4.2. The output of this network is a linear combination of the basis functions computed by the hidden layer nodes.
Figure 4.2 A Radial Basis Function Neural Network
4.2.2 RBF Training Algorithms
The RBF neural network can be used for … [Hus93][Tou1974].
4.3 Summary
Start here with the summary
5 PUT A TITLE HERE RELATED TO IMPLEMENTATION
This chapter discusses different aspects of the implementation. The first paragraph is a general introduction to the chapter and has no section number.
5.1 Introduction
This chapter describes the implementation details the single-digit recognition system illustrated in Figure 5.1. The recognition system consists of the following main components:
Figure 5.1 A Block Diagram of the Proposed System
Figure 5.2 Endpoints of the Digit “five”.
5.2 Title
5.2.1 This stage
To extract energy parameters of a Mel-frequency scale filter bank, computing a 512-point Fast Fourier transform (FFT) of each analysis frame simulates the filters.
(5.7)
The indices kL(p) and kH(p) correspond to lower and upper FFT coefficients of the band p as listed in Table 5.1.
Table 5.1 Mel-scale Frequency Bands and the Corresponding FFT Coefficients
|
Band p |
Center Freq. |
kL |
kH |
Band p |
Center Freq. |
kL |
kH |
|
0 |
100 |
2 |
3 |
10 |
1149 |
28 |
31 |
|
1 |
200 |
4 |
6 |
11 |
1320 |
32 |
36 |
|
2 |
300 |
7 |
8 |
12 |
1516 |
37 |
41 |
|
3 |
400 |
9 |
11 |
13 |
1741 |
42 |
47 |
5.2.2 Sub Section Title
Note the use of tables as well as figures is important to get information across to the reader. Table 5.2 shows…
Table 5.2 Phonetic Decomposition of Digits [Rab1993, Table 2.3]
|
Digit |
Phonemes |
# of Phonemes |
|
One |
W-AH-N |
3 |
|
Two |
T-UW |
2 |
|
Three |
TH-R-IY |
3 |
5.3 Summary
The previous sections of this chapter include …
6 PUT IN A TITLE RELATED TO EXPERIMENTS AND RESULTS
This chapter describes the experiments
6.1 Experiment One
Start here…
6.2 Experiment Two
Start here
6.2.1 Evaluation Set
Start here. Do not allow tables or figures to be split across pages.
|
ID |
Dlct. |
Gdr. |
Set |
Evl. |
Trn. |
Tst. |
ID |
Dlct. |
Gdr. |
Set |
Evl. |
Trn. |
Tst. |
|
ai |
P |
W |
0,2 |
√ |
√ |
|
ap |
P |
W |
1,2 |
|
√ |
√ |
|
al |
B |
M |
2 |
|
√ |
|
bn |
R |
M |
1,2 |
|
√ |
√ |
|
bi |
B |
W |
2 |
|
√ |
|
bw |
R |
W |
1,2 |
|
√ |
√ |
|
cb |
H |
M |
2 |
|
√ |
|
cd |
B |
W |
2 |
|
|
√ |
|
cl |
H |
W |
2 |
|
√ |
|
dr |
H |
W |
2 |
|
|
√ |
|
cr |
H |
M |
2 |
|
√ |
|
dw |
B |
W |
2 |
|
|
√ |
|
dc |
B |
W |
2 |
|
√ |
|
ep |
R |
W |
1,2 |
|
√ |
√ |
|
ea |
H |
W |
2 |
|
√ |
|
et |
P |
M |
1,2 |
|
√ |
√ |
6.2.2 Recognition Set #1
The …
6.2.3 Recognition Set #2
The …
6.2.4 Boundaries Detection
The …
Figure 6.1 Examples of Extra Subwords
6.3 Results Of Recognition Set #1
The results of this set of experiments are listed in Table 6.8.
Table 6.2 Average Recognition Rates in Manual and Automatic Systems
|
Size of |
Recognition % |
|
|
Training Set |
Manual System |
Automatic System |
|
2 |
98.48 |
60.10 |
|
3 |
97.84 |
73.05 |
|
4 |
99.07 |
79.02 |
|
5 |
99.11 |
82.74 |
|
8 |
99.44 |
87.60 |
6.4 Summary
This chapter has an…
7 CONCLUSIONS AND RECOMMENDATIONS
There are at least two sections here that are absolutely required. The preamble sentence is not required.
7.1 Conclusions
The results of experiments performed and the conclusions made for the work done in this seminar.
7.2 Recommendations
REFERENCES
[These are absolutely required – there must be archived papers referenced – these must be referred to in the text. url references must have the date of your last access.]
[Alg1993] V.R. Algazi, K.L. Brown, M.J. Ready, D.H. Irvine, C.L. Cadwell, S. Chung, “Transform Representation of the Spectra of Acoustic Speech Segments with Applications-I: General Approach and Application to Speech Recognition”, IEEE Transactions on Speech and Audio Processing 1(2):180-195, April 1993.
[Che1991] S. Chen, C.F.N. Cowan, P.M. Grant, “Orthogonal Least Squares Learning Algorithm for Radial Basis Function Networks”, IEEE Transactions on Neural Networks, 2(2):302-309, March 1991.
[Cis2000] Cisco Online, “Understanding Frame Relay Subinterfaces”, http//www.cisco.com/warp/public/779/smbiz/service/knowledge/wan/subifs.htm, Feb. 27, 2000.
[Gup1991] V.N. Gupta, M. Lennig, P. Mermelstein, P. Kenny, F. Seitz, D. O’Shaughnessy, “Using Phoneme Duration and Energy Contour Information to Improve Large Vocabulary Isolated-Word Recognition”, Proceedings ICASSP’91, 341-344, 1991.
[Hus1993] D.R. Hush, B.G. Horne, “Progress in Supervised Neural Networks: What’s New Since Lippmann?”, IEEE Signal Processing Magazine, 10(1):8-39, January 1993.
[Lip1987] R.P. Lippmann, “An Introduction to Computing with Neural Nets”, IEEE ASSP Magazine, 4(2):4-22, April 1987.
[Mar1995] S.M. Marcus, A.K. Syrdal, “Speech: Articulatory, Linguistic, Acoustic, and Perceptual Descriptions”, in Applied Speech Technology, ed. A. Syrdal, R. Bennett, S. Greenspan, CRC Press, Boca Raton, Florida, 1995.
[Pic1993] J.W. Picone, “Signal Modeling Techniques in Speech Recognition”, Proceedings of the IEEE, 81(9):1215-1247, September 1993.
[Rab1993] L.R. Rabiner, B.H. Juang, Fundamentals of Speech Recognition, Prentice-Hall, Englewood Cliffs, NJ, 1993.
[Rab1978] L.R. Rabiner, R.W. Schafer, Digital Processing of Speech Signals, Prentice-Hall, Englewood Cliffs, NJ, 1978.
[Sve1989] T. Svendsen, K.K. Paliwal, E. Harborg, P.O. Husoy, “An Improved Sub-Word Based Speech Recognizer”, Proceedings ICASSP’89, 108-111, May 1989.
[Tos1996] C. Tosuner, Single Spoken Digit Recognition Using RBF Networks, Unpublished Master’s Thesis, Technical University of Nova Scotia, Halifax, Nova Scotia, 1996.
[Tou1974] J.T. Tou, R.C. Gonzales, Pattern Recognition Principles, Addison-Wesley, MA, 1974.
BIBLIOGRAPHY
[Use the same format as in the REFERENCES section, however the entries in the BIBLIOGRAPHY are not referenced in the body of the text. So think of these as additional reading titles - something you read for background and which a reader of the report might find interesting.]