Assessment 3 is an individual assessment. The purpose of this assignment is to extend your knowledge of the concepts covered in this unit about different facets of AI. You are required to write and execute python code for the given tasks. You are also required to write a report which will have python code, output screenshots showing the answers to the questions, and an analysis of the generated outputs.

Please note that ALL submitted Assessment 3 reports are passed through a computerized copy detection system and it is extremely easy for the teaching staff to identify copied or otherwise plagiarised work.

  • Copying (plagiarism) can incur penalties ranging from deduction of marks to failing the course or even exclusion from the course or even exclusion from the University.
  • Please ensure you are familiar with the Academic Misconduct Procedures.

Task Description

 Task: Build AI (deep learning) based Image classification models

 Objective: Construct precise image classification model to determine from chest X-ray images whether or not a person has pneumonia.

Description: The normal chest X-ray depicts clear lungs without any areas of abnormal opacification in the image. Bacterial pneumonia typically exhibits a focal lobar consolidation whereas viral pneumonia manifests with a more diffuse interstitial pattern in both lungs. Now a days, Artificial Intelligence techniques are widely used to identity pneumonia from chest X-ray images. Since it is a matter of life, accuracy is essential for such AI based image classification model. In this task, you need to use Kaggle chest X-ray images dataset and build deep learning) based image classification models. This task will introduce you to essential techniques in image processing and Artificial intelligence.

You are required to divide AI-based image classification task into the following subtasks:

  • Data Collection and exploration: Download the chest X-ray dataset from Kaggle and save it in your local directory.
  • Pre-processing: Perform necessary pre-processing steps on the dataset, such as resizing images, converting grayscale to 3 channels, and normalization.
  • Model development: Develop two variant of deep learning image classification models to distinguish normal and pneumonia images from the given dataset. Split the dataset into training, validation and testing sets and train the model on the training set.
  • Evaluation: Evaluate the performance of the classification model on the testing set.

Report: Write a report which will have python code, output screenshots and an analysis of the generated outputs addressing all of the above subtasks.


Submit the following files on Moodle only:

  1. The answer template, called [StudentID]-report.docx. Replace [StudentID] with your actual student (Don’t submit report in a zip format)
  1. Python source code [StudentID].ipynb with a zip format containing python code, and any additional file needed to run the code.

Marking Criteria

 Assessment 3 will be marked based on the following criteria.

Data collection and exploration 5 marks
Pre-processing 10 marks
Model Development 18 marks
Evaluate the performance of the classifier Models 5 marks
Well-Written report with reflections 7 marks
Total 45 marks


To help you communicate, forums have been set up for you on the unit Moodle website for Assessments 1, 2 and 3. Please use them to help you work through your report.

Alternatively, you can take help from Academic Learning Centres (ALC). We have advisers situated in Academic Learning Centres at many CQUniversity locations, who offer generic group sessions, unit specific workshops, individual appointments, drop in centres, and print and online resources. ALC provides a range of services which include:

  • workshops online and on-campus (see our Moodle website for details)
  • online review of assignments
  • online query
  • one-on-one appointments in person, over the phone or online via Zoom