Fast image reconstruction in fluorescence optical tomography using data compression.

Area of Science:

• Biomedical Optics
• Image Reconstruction
• Computational Imaging

Background:

• Fluorescence optical tomography (FOT) generates large datasets (>10^7 samples) from multi-position CCD cameras.
• Storing the full system Jacobian is infeasible, often necessitating data subsampling.
• Subsampling can lead to loss of information and suboptimal reconstruction accuracy.

Purpose of the Study:

• To present a novel method for accelerating image reconstruction in FOT.
• To enable the use of full, large datasets without infeasible storage requirements.
• To improve the speed and potentially the accuracy of FOT reconstructions.

Main Methods:

• Image compression techniques are employed to manage large datasets.
• Explicit construction of a reduced-size system Jacobian is performed.
• Optimal inversion methods are applied to the compressed data and Jacobian.

Main Results:

• The method allows for the utilization of the complete dataset, avoiding subsampling.
• Explicit Jacobian construction results in a computationally manageable size.
• The approach leads to significantly faster image reconstruction times.

Conclusions:

• This method offers a viable solution for fast FOT reconstruction with very large datasets.
• It overcomes the storage and computational challenges associated with full data acquisition.
• The technique paves the way for more efficient and accurate biomedical imaging using FOT.