Imagine you’re trying to create a digital map of the brain using biomedical images. One of the challenges you face is aligning different samples into one consistent spatial coordinate system. This is where three dimensional deformable image registration (DIR) comes into play. Think of DIR as the method that can model the local non-linear deformation and align the anatomical structures in these images. However, there’s a catch. Existing DIR methods struggle with large non-linear deformations that are not handled well by traditional approaches. Enter the progressive image registration strategy based on deep self-calibration. This approach tackles the problem head-on, without causing any loss of image details or introducing additional parameters. Not only that, but it also employs a hierarchical registration strategy for accurate multi-scale progressive registration. And, to top it all off, this method even implements dynamic dataset augmentation for improved results! The researchers evaluated their proposed method and achieved promising results, outperforming other state-of-the-art approaches for deformable image registration.

Three dimensional deformable image registration (DIR) is a key enabling technique in building digital neuronal atlases of the brain, which can model the local non-linear deformation between a pair of biomedical images and align the anatomical structures of different samples into one spatial coordinate system. And thus, the DIR is always conducted following a preprocessing of global linear registration to remove the large global deformations. However, imperfect preprocessing may leave some large non-linear deformations that cannot be handled well by existing DIR methods. The recently proposed cascaded registration network gives a primary solution to deal with such large non-linear deformations, but still suffers from loss of image details caused by continuous interpolation (information loss problem). In this article, a progressive image registration strategy based on deep self-calibration is proposed to deal with the large non-linear deformations without causing information loss and introducing additional parameters. More importantly, we also propose a novel hierarchical registration strategy to quickly achieve accurate multi-scale progressive registration. In addition, our method can implicitly and reasonably implement dynamic dataset augmentation. We have evaluated the proposed method on both optical and MRI image datasets with obtaining promising results, which demonstrate the superior performance of the proposed method over several other state-of-the-art approaches for deformable image registration.

Read Full Article (External Site)