Accurate and automated retinal vessel segmentation is crucial for the diagnosis and prevention of many diseases. However, current methods still fall short in processing vessels of different scales, especially in the case of fine vessels. To solve the above issues, this paper proposes a multiscale deep fusion network for retinal vessel segmentation. Specifically, this paper first designs a Frangi enhancement module. The FE module leverages the output response from the Frangi filter in combination with the feature map of the neural network, which assists the network in segmenting more fine blood vessels. Secondly, this paper proposes a multiscale bidirectional fusion (MBF) module. The MBF module comprises dilated convolutions with different dilation rates and bidirectional attention mechanisms, designed to extract vascular features at various scales and establish connections between distant pixels. Finally, this paper investigates a dual-threshold decision (DTD) algorithm, which is used to solve the problem of ambiguous pixel determination near the single threshold. The proposed MDF-Net is evaluated on three typical fundus image datasets (DRIVE, STARE, CHASE-DB1), and the experimental results indicate that MDF-Net exhibits superior performance compared to other state-of-the-art methods. Additionally, the application of the DTD algorithm significantly improves the SE (0.8615, 0.8804, 0.8792) metric, effectively enhancing the connectivity of the vessels.
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.