Efficient Prealignment of CT Scans for Registration through a Bodypart Regressor
This work addresses the need for efficient prealignment in medical imaging registration, particularly for CT scans, but it appears incremental as it builds on an underexplored idea without major breakthroughs.
The paper tackled the problem of aligning CT scans from different timepoints by using a CNN as a bodypart regressor to output relative height, enabling fast and robust prealignment compared to iterative methods, with preliminary results confirming its efficiency.
Convolutional neural networks have not only been applied for classification of voxels, objects, or images, for instance, but have also been proposed as a bodypart regressor. We pick up this underexplored idea and evaluate its value for registration: A CNN is trained to output the relative height within the human body in axial CT scans, and the resulting scores are used for quick alignment between different timepoints. Preliminary results confirm that this allows both fast and robust prealignment compared with iterative approaches.