Dep2SMS: Simultaneous Multi-Slice Reconstruction via Deep Learning with Auxiliary Depth Camera Guidance

Motivation: Simultaneous Multi-Slice (SMS) reconstruction achieves slice separation by single-band calibration data or coil sensitivity maps through pre-scanning, resulting in SMS inefficiency. Goal(s): Achieve real-time and robust contour acquisition through the depth camera. Proposed a SMS reconstruction method via deep learning with auxiliary depth camera guidance. Approach: We obtain the contour by placing the depth camera, locating phantom, and fixed markers within the MRI. We propose a Mamba-based network based on the auxiliary depth camera guidance to achieve faithful SMS reconstruction. Results: The contour captured by the depth camera demonstrates effectiveness with less time-consuming and more robust, and the proposed Dep2SMS achieves outstanding reconstruction. Impact: We novelty introduce the depth camera into the MRI system to capture contour to assist SMS reconstruction. We utilize the Mamba-based framework with intra-patch convolution and linear-complexity long-range attention for SMS reconstruction to capture fine structural and global texture features.