Dual Branch Prior-SegNet: CNN for Interventional CBCT using Planning Scan and Auxiliary Segmentation Loss
This work addresses image quality in medical imaging for interventional procedures, but it is incremental as it builds upon prior methods.
The paper tackles sparse view interventional CBCT reconstruction by incorporating a high-quality planning scan and an auxiliary segmentation loss for instruments, achieving a performance gain of over 2.8 dB PSNR and robustness to rotations up to ±5.5 degrees.
This paper proposes an extension to the Dual Branch Prior-Net for sparse view interventional CBCT reconstruction incorporating a high quality planning scan. An additional head learns to segment interventional instruments and thus guides the reconstruction task. The prior scans are misaligned by up to +-5deg in-plane during training. Experiments show that the proposed model, Dual Branch Prior-SegNet, significantly outperforms any other evaluated model by >2.8dB PSNR. It also stays robust wrt. rotations of up to +-5.5deg.