Answers are grounded in this paper and cite their sources.
There are various medical imaging methods which have been used broadly in clinical and medical research. Consequently, the interests in registering and finding similarities of different images for diagnosis, treatment, and the sake of basic science are increasing. As images are typically captured at different times, angles, and often by different modalities, registering (or aligning) one image with another is challenging. In general, the accuracy of registration techniques will affect the performance and robustness of all subsequent analysis. We propose an efficient 3D medical image registration method based on sparse coding and belief propagation for Computed Tomography (CT) and Magnetic Resonance (MR) imaging. We used 3D image blocks as the input features and then we employed sparse coding with a dictionary of the features to find a set of the candidate voxels. To select optimum matches, belief propagation was subsequently applied on a factor graph of voxels generated by these candidate voxels. The outcome of belief propagation was interpreted as a probabilistic map of aligning the candidate voxels to the source voxels. We compared our proposed method (3D-SCoBeP) with the state-of-the-art medical image registration, MIRT and GP-Registration algorithm. Our objective results based on Root Mean Square Error (RMSE) are smaller than those from MIRT and GP-Registration. Our results prove the effectiveness of our algorithm in registering the reference image to the source image.
No takes yet. Share an insight, caveat, or question.
Roozgard et al. (2014) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: