Primary supervisor

Jesmin Nahar

1. Background

Skin cancer, including melanoma, basal cell carcinoma (BCC), and squamous cell carcinoma (SCC), is a significant health concern in Australia. Early detection and accurate risk assessment can support timely clinical assessment and treatment.

2. Research Aim

This research aims to develop an AI and machine-learning framework for early detection, risk classification, clinical decision support, and forecasting of skin cancer.

3. Data and Methods

The study will investigate clinical, demographic and/or skin-lesion image data. Machine-learning and deep-learning techniques, such as Logistic Regression, Random Forest, Support Vector Machine, XGBoost/LightGBM and deep-learning models, will be evaluated.

4. Prediction and Risk Classification

The models will predict the likelihood of skin-cancer risk and classify lesions or patients into appropriate risk categories. Model performance will be assessed using accuracy, sensitivity, specificity, precision, recall, F1-score and ROC-AUC.

5. Explainable AI and Decision Support

Explainable AI (XAI) techniques will be used to identify important factors contributing to model predictions. The resulting risk information could support clinicians in identifying potentially high-risk cases requiring further assessment.

6. Forecasting

Where suitable longitudinal or population-level data are available, machine-learning and time-series approaches will be explored to forecast future skin-cancer trends, risk patterns, or healthcare demand.

7. Expected Outcome

The research aims to develop an interpretable and reliable AI-enabled decision-support framework for skin-cancer risk assessment and early detection. The system will be designed to support clinical decision-making rather than replace professional diagnosis, with emphasis on accuracy, explainability, robustness and responsible AI.

Aim/outline

Aim

To investigate the application of Artificial Intelligence (AI) and Machine Learning (ML) for the early detection, risk classification and prediction of skin cancer, with the aim of supporting timely and informed clinical decision-making.

Outline

The research will:

  • Analyse clinical, demographic and/or skin-lesion image data.
  • Develop and compare machine-learning and deep-learning models for skin-cancer prediction and risk classification.
  • Evaluate model performance using appropriate statistical and clinical metrics.
  • Apply Explainable AI (XAI) to improve the interpretability of predictions.
  • Investigate the potential of AI for clinical decision support.
  • Where suitable longitudinal data are available, explore forecasting of skin-cancer trends and healthcare demand.
  • Assess model robustness, generalisability and responsible use in healthcare.

Required knowledge

Required Knowledge

Students should have a basic to intermediate understanding of:

  • Artificial Intelligence and Machine Learning
  • Python and/or R programming
  • Data preprocessing and feature engineering
  • Supervised machine-learning techniques, including classification and prediction
  • Model evaluation and performance metrics
  • Basic statistics and data analysis
  • Basic understanding of deep learning/computer vision is desirable but not essential
  • An interest in healthcare applications of AI and responsible/explainable AI