Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

EENGM0004: Engineering Research Skills

IPP Writing

Async 2: Project Planning

Key Elements of Project Planning

Background: context, impact, importance and urgency.

Statement of the problem: state of the arts, existing solutions, TRL

Aims and objectives: general purpose vs measurable outcomes.

Methodology: workpackages, tasks, activities, objective/approach/deliverable

Work plan: PERT/Gantt charts, milestones

Risk management: identification, scales and mitigation

Interim Project Planning (IPP) Report

To present an outline understanding of your project

Background

Context

State of the arts

Importance and impact

To frame or define your project

Research gaps

Content of research

Aims and objectives

General approaches

The Recommended Structure of an IPP Report

1. Background

2. Aims and Objectives

3. Related Work

4. Conclusion

N.B. Templates have been provided on Blackboard for IPP writing (in both MS Word and Latex formats).

Background

Background of this research topic

What is the context: e.g., new technology, climate change and commercial applications.

Why this work is important?

What is the current problem? Why has it not been solved?

What benefits can be gained by solving this problem?

General guidance

Logical statements with proper reference (2-3 paragraphs).

It should connect well with the aims and objectives of the project.

Don’t confuse Background with Related Work.

Background: An Example

Project title

Deep learning based video compression

Background

The demand for greater quantities of higher quality, more immersive video content is the primary driver for internet, broadcasting and surveillance technologies. CISCO predict that, by 2022, there will be 4.8ZB of global internet traffic per year with 82% being video [1]. In this context, how we represent and communicate video (via compression) is key in ensuring that the content is delivered at an appropriate quality, while maintaining compatibility with the transmission bandwidth. Although a series of conventional coding standards [2-5] have been developed in the past few decades, these compression algorithms are still built within the rate-distortion optimisation framework as their predecessors, and neither of them is anticipated to fully meet the growing demand for future media consumption [6]. A step change in coding performance, going well beyond classical signal compression, is therefore urgently required to ease the tension between the available transmission bandwidth and the high bit rates demanded by video content.

Exercise: Background

Write a draft of the background section for your project

Think about your project based on the title and description (or your first meeting).

Write down a few bullet points for each sentence.

Find a couple of key references to support your statements.

Discuss with your classmate who is setting next to you

Provide comments and suggestions.

Improve your draft.

Aims and Objectives

What do you plan to achieve? – Discuss with your project supervisors.

Aims: the general purpose of the project - the big picture!

What problem the project aims to solve?

What kind of innovation does it want to achieve?

Objectives: measurable outcomes of project activities

Specific targets the project will achieve.

How do the readers measure whether you have achieved the objectives? Try to be quantitative.

When defining objectives, think about how easy/difficult they can be achieved.

General guidance

1-2 paragraphs with bullet points.

Don’t confuse project activities with aims and objectives!

Aims and Objectives: An Example

Project title

Deep learning based video compression

Aims and Objectives

Inspired by recent breakthroughs in AI technology, the aim of this project is to develop a new deep learning based video compression framework, which is expected to make a step change in coding gain, targeting rate-quality performance improvement of more than 30% over the state of the art – a typical improvement figure expected from a new generation of coding standard. To achieve this, the objectives are:

1. To create diverse databases for training and evaluating deep video coding tools.

2. To generate baseline results for benchmarking using conventional coding methods.

3. To develop and train new CNN-based coding tools with evident coding gains over baseline methods.

Aims and Objectives - Problems (1)

Project: Perceptually-inspired video coding

Aims and Objectives (draft): This project aim is to develop new coding algorithms which exploit the perceptual redundancy of video signals. The objectives are to design new algorithms using Matlab and C++, and evaluate their performance based on objective quality metrics, such as VMAF, and subjective test.

Aims and Objectives - Problems (1)

Project: Perceptually-inspired video coding

Aims and Objectives (draft): This project aim is to develop new coding algorithms which exploit the perceptual redundancy of video signals. The objectives are to design new algorithms using Matlab and C++, and evaluate their performance based on objective quality metrics, such as VMAF, and subjective test.

Analysis

The aim is well presented, with clear description on the final target.

The objectives are not designed properly, with focuses on detailed approaches. Bullet points should be used.

Aims and Objectives - Problems (1)

Project: Perceptually-inspired video coding

Aims and Objectives (draft): This project aim is to develop new coding algorithms which exploit the perceptual redundancy of video signals. The objectives are to design new algorithms using Matlab and C++, and evaluate their performance based on objective quality metrics, such as VMAF, and subjective test.

Analysis

The aim is well presented, with clear description on the final target.

The objectives are not designed properly, with focuses on detailed approaches. Bullet points should be used.

Revision

Aims and Objectives: This project aim is to develop new coding algorithms which exploit the perceptual redundancy of video signals. To achieve this aim, the following objectives need to be fulfilled.

To develop new perceptual quality metrics which correlate well with subjective opinions.

To design novel perceptually-inspired coding tools with improved rate quality performance.

To generate the evaluation results of the new coding tools benchmarked against conventional methods.

Aims and Objectives - Problems (2)

Project: Deep learning based Image Denoising

Aims and Objectives (draft): This project is aiming to solve the problem of image denoising. The objectives are:

To study the construction of feed forward denoising convolutional neural network (dnCNNs).

To apply the structure, learning algorithm and regularisation method for image denoising.

Aims and Objectives - Problems (2)

Project: Deep learning based Image Denoising

Aims and Objectives (draft): This project is aiming to solve the problem of image denoising. The objectives are:

To study the construction of feed forward denoising convolutional neural network (dnCNNs).

To apply the structure, learning algorithm and regularisation method for image denoising.

Issues

The aim is too general. It does not reflect the target of the research project.

The objectives are not well presented either. It is not clear what to achieve at each stage of the project.

Aims and Objectives - Problems (2)

Project: Deep learning based Image Denoising

Aims and Objectives (draft): This project is aiming to solve the problem of image denoising. The objectives are:

To study the construction of feed forward denoising convolutional neural network (dnCNNs).

To apply the structure, learning algorithm and regularisation method for image denoising.

Issues

The aim is too general. It does not reflect the target of the research project.

The objectives are not well presented either. It is not clear what to achieve at each stage of the project.

Revision

Aim: This project aims to develop new image denoising algorithms based on deep learning techniques, which are expected to offer significant performance gains over classic denoising approaches. Objectives:

To collect diverse training material for optimising CNN-based denoising models.

To design a new CNN architecture with improved denoising performance.

To generate denoising results using the proposed method compared to existing denoising algorithms.

Exercise: Aims and Objectives

Summarise the aims and objectives of your project

Think about your project based on the title and description (or your first meeting).

Write down the aims and objectives (draft).

Discuss with your classmate who is setting next to you

Provide comments and suggestions.

Improve your draft.

Related Work

The state of the arts in this research area (Critical Review)

What has been done so far: facts.

How much of the problem has been solved, how much remains to be solved?

What are the good and not-so-satisfactory aspects of their methodologies?

What is your reason of saying so (evidence): e.g. results and references.

General guidance

Form viewpoints first before writing.

This should lead to what you are going to do.

0.5-1 page with references in the end.

Related Work: An Example

Project title

Deep learning based video compression

Related Work (one example paragraph)

Existing deep learning-based picture coding algorithms can be classified into two primary categories [1]. The first relates to end-to-end training and optimisation using auto-encoder type architectures. (...descriptions on typical works in this category...) Although the solutions in this category are not yet competitive with the latest standardised codecs, such as VVC and AV1, they demonstrate significant potential for the future [2-3]. A second class contains algorithms that are designed to enhance individual coding tools within a standard codec configuration. (...descriptions on typical works in this category...) However, it is noted that these approaches are typically associated with much higher computational complexity, which may lead to practical issues when employed for real time applications.

Conclusion

What should be included?

Summarise the report and highlight the research gaps that you have identified.

Inlcude some preliminary ideas on how you will conduct your research work.

▶ Which methods from the literature are most likely to give you expected results in your project?

▶ Is the choice clear, or is there a need to compare tools/approaches as part of your project (i.e. it becomes an objective in itself)?

Justify the urgency and importance to do your project and the proposed methodology.

▶ Have you identified questions that need to be addressed in your project?

▶ What is the difference that your project will bring in compared to the research review you have just completed?

General guidance

The Conclusion session should be 2-3 paragraphs long.

NOTE: The guidance provided above for the IPP writing is just for recommendation! You should follow your supervisor’s advice specifically for your project.

References

Types of Bibliography Styles

MLA: Girod, Bernd, et al. “Distributed video coding.” Proceedings of the IEEE 93.1 (2005): 71-83.

APA: Girod, B., Aaron, A. M., Rane, S., & Rebollo-Monedero, D. (2005). Distributed video coding. Proceedings of the IEEE, 93(1), 71-83.

Chicago: Girod, Bernd, Anne Margot Aaron, Shantanu Rane, and David Rebollo-Monedero. “Distributed video coding.” Proceedings of the IEEE 93, no. 1 (2005): 71-83.

Harvard: Girod, B., Aaron, A.M., Rane, S. and Rebollo-Monedero, D., 2005. Distributed video coding. Proceedings of the IEEE, 93(1), pp.71-83.

Vancouver: Girod B, Aaron AM, Rane S, Rebollo-Monedero D. Distributed video coding. Proceedings of the IEEE. 2005 Jun 27;93(1):71-83.

General Guidance

The bibliography style should be consistent in a document.

Various reference management software can be used for different editing environment: e.g., EndNote for MS Word and JabRef for Latex.