Single-view cone-beam x-ray luminescence optical tomography based on Group_YALL1 method

Abstract

Single-view cone beam x-ray luminescence optical tomography (CB-XLOT) has the merit of short data acquisition time, which is important for resolving fast biological processes in vivo. However, challenges remain in the reconstruction of single-view CB-XLOT. In our previous work, by using the sparsity-based reconstruction method, we have demonstrated the feasibility of single-view CB-XLOT. But, when the imaging conditions become complicated (e.g. multiple adjacent nanophosphors (NPs) contained in imaged object), it is difficult to resolve each NP by the previous method. To solve the problem, we hereby present a sparsity reconstruction method based on group information, termed Group_YALL1. The imaging performance of single-view CB-XLOT can be further improved by utilizing the group sparsity characteristic of NPs as a priori knowledge of reconstruction constraint. To assess the capability of the method, we used a customized CB-XLOT/XCT system to perform the numerical simulation and physical phantom experiments. The experimental results demonstrate that compared with the former sparse reconstruction method (e.g. YALL1), the proposed Group_YALL1 method can accurately resolve the NPs embedded in the object, even if they are close to each other. The acquired location error is less than 1 mm. Hence, this method has the potential to greatly reduce the data acquisition time while preserving a high imaging quality.