Deformable image registration is crucial for medical image analysis, yet the complexity of deep learning networks often limits their deployment on resource-limited devices. Current distillation methods in registration tasks fail to effectively transfer complex deformation handling capabilities to non-iterative lightweight networks, leading to insignificant performance improvement. To address this, we propose the Framework for Iterative to Non-iterative Distillation (FIND), which efficiently transfers these capabilities to a Non-Iterative Lightweight (NIL) network. FIND employs a dual-step process: first, using recurrent distillation to derive a high-performance non-iterative teacher assistant from an iterative network; second, using advanced feature distillation from the assistant to the lightweight network. This enables NIL to perform rapid, effective registration on resource-limited devices. Experiments across four datasets show that NIL can achieve up to 60 times faster performance on CPU and 89 times on GPU than compared deep learning methods, with superior registration accuracy improvements of up to 3.5 points in Dice scores.

In clinical practice, it is desirable for medical image segmentation models to be able to continually learn on a sequential data stream from multiple sites, rather than a consolidated dataset, due to storage cost and privacy restrictions. However, when learning on a new site, existing methods struggle with a weak memorizability for previous sites with complex shape and semantic information, and a poor explainability for the memory consolidation process. In this work, we propose a novel Shape and Semantics-based Selective Regularization ( [Formula: see text]) method for explainable cross-site continual segmentation to maintain both shape and semantic knowledge of previously learned sites. Specifically, [Formula: see text] method adopts a selective regularization scheme to penalize changes of parameters with high Joint Shape and Semantics-based Importance (JSSI) weights, which are estimated based on the parameter sensitivity to shape properties and reliable semantics of the segmentation object. This helps to prevent the related shape and semantic knowledge from being forgotten. Moreover, we propose an Importance Activation Mapping (IAM) method for memory interpretation, which indicates the spatial support for important parameters to visualize the memorized content. We have extensively evaluated our method on prostate segmentation and optic cup and disc segmentation tasks. Our method outperforms other comparison methods in reducing model forgetting and increasing explainability. Our code is available at https://github.com/jingyzhang/S3R.

Percutaneous coronary intervention is widely applied for the treatment of coronary artery disease under the guidance of X-ray coronary angiography (XCA) image. However, the projective nature of XCA causes the loss of 3D structural information, which hinders the intervention. This issue can be addressed by the deformable 3D/2D coronary artery registration technique, which fuses the pre-operative computed tomography angiography volume with the intra-operative XCA image. In this study, we propose a deep learning-based neural network for this task. The registration is conducted in a segment-by-segment manner. For each vessel segment pair, the centerlines that preserve topological information are decomposed into an origin tensor and a spherical coordinate shape tensor as network input through independent branches. Features of different modalities are fused and processed for predicting angular deflections, which is a special type of deformation field implying motion and length preservation constraints for vessel segments. The proposed method achieves an average error of 1.13 mm on the clinical dataset, which shows the potential to be applied in clinical practice.

Vessel enhancement is an important pre-processing step of applications in vessel image analysis. However, most of the current methods are developed merely based on the intensity variety inside and outside vessel instead of considering the vessel path, which emphasizes the vascular structures via characterizing additional connectivity and length information. Aiming at further utilizing beneficial length information of vessels, we propose a novel method to impose length constraint on Hessian information for vessel enhancement. Specifically, Eigen analysis of multiscale Hessian matrix has been taken at each pixel for the local vesselness response and direction information. Then, vessel path is searched along each pixel's direction, as well as maintains the property of curvilinear smoothness. The proposed method is compared with three conventional vessel enhancement methods. The experiment results show that our proposed approach has the advantages of the fine response of low-contrast vessel region and less noise background. In addition, the quantity evaluation indicates that a state-of-art vessel enhancement performance could be achieved compared with other methods.

Ultrasound video segmentation is a challenging task due to low contrast, shadow effects, complex noise statistics, and the need for high precision and efficiency in real time applications such as operation navigation and therapy planning. In this paper, we propose a wavelet frame based video segmentation framework incorporating different noise statistics and sequential distance shape priors. The proposed individual frame nonconvex segmentation model is solved by a proximal alternating minimization algorithm, and the convergence of the scheme is established based on the recently proposed Kurdyka--Łojasiewicz property. The performance of the overall method is demonstrated through numerical results on two real ultrasound video data sets. The proposed method is shown to achieve better results compared to the related level sets models and edge indicator shape priors, in terms of both segmentation quality and computational time.

We propose a navigation system for transcatheter aortic valve implantation that employs a magnetic tracking system (MTS) along with a dynamic aortic model and intra-operative ultrasound (US) images. This work is motivated by the desire of our cardiology and cardiac surgical colleagues to minimize or eliminate the use of radiation in the interventional suite or operating room. The dynamic 3-D aortic model is constructed from a preoperative 4-D computed tomography dataset that is animated in synchrony with the real time electrocardiograph input of patient, and then preoperative planning is performed to determine the target position of the aortic valve prosthesis. The contours of the aortic root are extracted automatically from short axis US images in real-time for registering the 2-D intra-operative US image to the preoperative dynamic aortic model. The augmented MTS guides the interventionist during positioning and deployment of the aortic valve prosthesis to the target. The results of the aortic root segmentation algorithm demonstrate an error of 0.92±0.85 mm with a computational time of 36.13±6.26 ms. The navigation approach was validated in porcine studies, yielding fiducial localization errors, target registration errors, deployment distance, and tilting errors of 3.02±0.39 mm, 3.31±1.55 mm, 3.23±0.94 mm, and 5.85±3.06(°) , respectively.

Determining liver segmentation accurately from MRIs is the primary and crucial step for any automated liver perfusion analysis, which provides important information about the blood supply to the liver. Although implicit contour extraction methods, such as level set methods (LSMs) and active contours, are often used to segment livers, the results are not always satisfactory due to the presence of artifacts and low-gradient response on the liver boundary. In this paper, we propose a multiple-initialization, multiple-step LSM to overcome the leakage and over-segmentation problems. The multiple-initialization curves are first evolved separately using the fast marching methods and LSMs, which are then combined with a convex hull algorithm to obtain a rough liver contour. Finally, the contour is evolved again using global level set smoothing to determine a precise liver boundary. Experimental results on 12 abdominal MRI series showed that the proposed approach obtained better liver segmentation results, so that a refined liver perfusion curve without respiration affection can be obtained by using a modified chamfer matching algorithm and the perfusion curve is evaluated by radiologists.

In this paper, we present a novel hybrid deformation model using global mass-spring medial representation structures and local finite element model. We employ the hybrid models, by fully calculating the FEM deformation in the local operation part while only calculating the global deformation by medial representation method. To achieve the real-time requirement of realistic deformable modeling, it is necessary to use the GPU parallel computing for FEM on regional deformation details, so the major calculation work in the conjugate gradient solver for the solution matrix is moved from CPU to GPU to accelerate the effectiveness. Evaluation and experiments are also discussed.

An iterative classification method developed for 2D CT head data classification problem and using both statistical and spatial information is introduced in this paper. The method reduces the chance of misclassification, preserving the contiguity of tissue classes. This method is minimally supervised so that it enforces a relation between tissues and classes. In later iterations high-confidence points are used to help classify nearby ambiguous points, based on the assumption that points in close proximity and of comparable intensities are probably representing the same tissue class.

We present a new, efficient and easy to use collision detection scheme for real-time collision detection between highly deformable tetrahedral models. Tetrahedral models are a common representation of volumetric meshes which are often used in physically based simulations, e.g. in Virtual surgery. In a deformable models environment collision detection usually is a performance bottleneck since the data structures used for efficient intersection tests need to be rebuilt or modified frequently. Our approach minimizes the time needed for building a collision detection data structure. We employ an infinite hierarchical spatial grid in which for each single tetrahedron in the scene a well fitting grid cell size is computed. A hash function is used to project occupied grid cells into a finite ID hash table. Only primitives mapped to the same hash index indicate a possible collision and need to be checked for intersections. This results in a high performance collision detection algorithm which does not depend on user defined parameters and thus flexibly adapts to any scene setup.

In this paper, we propose a novel multistage method for three-dimensional (3-D) segmentation of medical images and a new radial distance-based segmentation validation approach. For the 3-D segmentation method, we first employ a morphological recursive erosion operation to reduce the connectivity between the region of interest and its surrounding neighborhood; then we design a hybrid segmentation method to achieve an initial result. The hybrid approach integrates an improved fast marching method and a morphological reconstruction algorithm. Finally, a morphological recursive dilation is employed to recover any lost structure from the first stage of the multistage method. This approach is tested on 12 CT and 3 MRI images of the brain, heart, and kidney, to demonstrate the effectiveness and accuracy of this technique across a variety of imaging modalities and organ systems. In order to validate the multistage segmentation method, a novel radial distance-based validation method is proposed that uses a global accuracy (GA) measure. The GA is calculated based on local radial distance errors (LRDE), where LRDE are calculated on the radii emitted from points along the skeleton of the object rather than the centroid, in order to accommodate more complicated organ structures. The experimental results demonstrate that the proposed multistage segmentation method is fast and accurate, with comparable performance to existing segmentation methods, but with a significantly higher execution speed.

We describe a method for automatic recognition of abdominal organs such as kidneys, spleen, stomach, and liver from computerised tomography (CT) images using 3D mathematical morphology. Morphological approaches provide the theory and tools to analyze shapes directly. This characteristic enables analyzing and recognizing abdominal organs according to size and gray level features. Our system consists of extraction part and recognition part. Differential top-hats and conditional dilation are used for the extraction part. Also a combination of recursive erosion and geodesic influence is found to be effective for separating touched organs. Recognition is based on a simple likelihood decision with organ's position and size. We obtained a recognition rate of about 91% for nine organs of four CT images.

Current character extraction systems for scene images are not robust for most real-world applications. In contrast, the system present here achieves robust performance by using morphological segmentation. This paper describes a new morphological segmentation algorithm-differential top-hats (DTT). In, addition, a complete system for extraction of characters from color scene images is presented. The system was verified through experiments on sequences of outdoor color images with varying external conditions. A high average extraction rate of 95% is obtained.