Our lab investigates machine learning techniques to establish numerical observers and approximate the ideal observer for task-based image quality assessment. Few projects in this category are as follows:

• Machine learning-based numerical observer
  • Approximating the Ideal Observer and Hotelling Observer for binary signal detection tasks by use of supervised learning methods
  • Approximating the Ideal Observer for joint signal detection and estimation tasks
• Task-based image quality assessment for medical imaging processing techniques
  • Assessing the impact of deep neural network-based image denoising on binary signal detection tasks
  • Task-based evaluation of deep image super-resolution in medical imaging
• Learning stochastic object models from imaging measurements