Optimization of acquisition geometry for intra-operative tomographic imaging

Area of Science:

• Medical Imaging
• Computational Science

Background:

• Tomographic reconstruction typically requires densely sampled acquisition geometries for well-posed linear systems.
• Intra-operative setups with freehand-guided sensing face challenges in enforcing optimal sampling compared to gantry-based systems.

Purpose of the Study:

• To develop an incremental method for monitoring the numerical condition of tomographic reconstruction systems.
• To present an approach for optimizing detector positions in freehand-guided intra-operative imaging.

Main Methods:

• Utilized singular value decomposition (SVD) of the system matrix to assess numerical condition.
• Employed a randomized optimization scheme to identify optimal detector positions.
• Validated the approach through simulations of intra-operative functional imaging and robot-controlled phantom experiments.

Main Results:

• Demonstrated the feasibility of incrementally monitoring system matrix condition.
• Showcased an effective randomized optimization strategy for detector positioning.
• Confirmed the practical applicability in simulated and experimental intra-operative scenarios.

Conclusions:

• The proposed method effectively monitors and optimizes acquisition geometries for intra-operative tomographic reconstruction.
• Randomized optimization of detector positions can enhance the well-posedness of the linear system in freehand-guided imaging.
• This technique holds promise for improving image quality and reliability in real-time surgical guidance.