Geodesic voting method is known as a powerful tool for extracting curvilinear structures, which is able to find a tree structure from a single point. However, this method may fail to generate accurate results in complex scenarios such as complex network-like structures, due to the limitation of single source point. In order to solve this problem, we propose an adaptive curvature-penalized geodesic voting method where multiple source points with geometric voting constraint can be used for constructing the voting score map. In addition, we exploit the introduced adaptive geodesic voting method for the task of retinal vessel tracking, in conjunction with a deep learning-based junction points detection procedure. Experimental results on both synthetic images and retinal images prove the efficiency of the introduced adaptive geodesic voting method.
In this paper, we introduce a new multi-center instance segmentation model based on the deep learning technique, as a generalization of the classical polarmask model. In contrast to the original polarmask model which imposes a star-convexity shape to approximate the target region, we propose to establish a multi-center model which allows representing the target region via multiple star convex shapes. For this purpose, we extract a set of points, each of which is taken as the centers of star convex shapes, to compute multiple star convex shapes. As a consequence, the final segmentation mask can be naturally generated using the union of all of the detected star convex shapes. Experimental results show that the multi-center polarmask model can achieve more advanced performance on the COCO dataset. In addition, the introduced model provides the possibility for real-time applications.
It is a popular way to incorporate the active contour evolution scheme into the multiscale image decomposition and reconstruction procedure, so as to enhance the image segmentation accuracy. However, most of these models are carried out by the level set formulation which cost much computation time. In this paper, we propose a new image segmentation model that combines the circular geodesic model with an adaptive cut and the multiscale image processing. As a consequence, the proposed model can blend the benefits from both of the geodesic models and the multiscale image analysis method. Experimental results show that the proposed multiscale geodesic model indeed outperforms the circular geodesic model with an adaptive cut in solving the image segmentation problem in the presence of strong noise.
KEYWORDS: Image segmentation, Medical imaging, Network architectures, Liver, Education and training, Anisotropy, Prostate, Deep learning, 3D image processing
Finding contours of interest from medical images is an important task in the field of medical image analysis. The current deep learning-based image segmentation approaches have obtained promising results. However, most of these models do not take into account the anisotropy and asymmetric features which play an important role in describing the target contours. In order to address this issue, we propose new loss-function applied to the deep learning model with dense distance regression, which can benefit the edge-based features, thus able to improve the stability of the segmentation procedure and to reduce the probability of outliers in the segmentation results. The introduced loss function is embedded into the deep learning model, which can perform an end-to-end image segmentation procedure for medical images. Ablation experiments were done with other loss functions and three datasets were used to verify whether this loss function is effective. SOTA results were obtained for the proposed loss function in this paper compared to the recently designed method for reducing the boundary error.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.