Informed Adaptive Sensing.

Area of Science:

• Computational imaging
• Applied mathematics
• Machine learning

Background:

• Many inverse problems rely on sequentially acquired data.
• Existing methods for data acquisition can be inefficient.

Purpose of the Study:

• To develop an adaptive sensing method for solving inverse problems with sequentially acquired data.
• To improve measurement efficiency and solution accuracy in inverse problems.

Main Methods:

• Utilizing an estimate of mutual information from a generative model's empirical conditional distribution.
• Guiding sensor reconfiguration based on measurements acquired so far.
• Applying the method to magnetic resonance image reconstruction and other inverse problems.

Main Results:

• The informed adaptive sensing method demonstrated superior performance compared to random sampling, variance-based sampling, sparsity-based methods, and compressed sensing.
• The approach was validated on both synthetic and real-world datasets, with a focus on image data.
• The method showed applicability to a broader range of problems beyond image reconstruction.

Conclusions:

• The proposed adaptive sensing strategy offers a more efficient and effective approach to data acquisition for inverse problems.
• Leveraging generative models and mutual information enables intelligent, on-the-fly sensor reconfiguration.
• The method's ability to generalize using learned models like deep neural networks highlights its potential for broad application.