Abstract
Tensor fields can encode local image features for downstream algorithms. Common examples include the structure tensor and Hessian, which encode the local gradient and curvature. However, the results are sensitive to noise because they rely on finite difference methods. While this can be mitigated through low-pass filtering, this comes at the cost of resolution. Tensor voting (TV) provides a method of distributing tensor information across neighboring pixels based on Gestalt principles. However, these basic assumptions do not translate well to biomedical images, particularly surface and fiber extraction. In this paper, we add several features to the TV framework enabling more robust control of the refined field. Specifically, we reformulate TV to enable (1) tuning the vote field’s scope to reinforce highly directed structures, (2) normalized energy distribution so that TV can be applied iteratively or with other algorithms, (3) a closed-form solution for all terms and the corresponding plate tensor, and (4) a GPUbased algorithm for fast processing of large images.
Visualization
This is a demo of the 3D visualization toolkit for tensor fields. While the paper focuses on 2D tensor voting, this implementation extends to 3D and is available in the same repository. The toolkit enables interactive exploration of tensor fields, helping to analyze their structure and behavior.
Resources
Paper: Our paper is available on IEEE and locally.
Poster: Our poster is available here.
Code: Our code is available on Github.
Data: Our data (Serial EM) is available upon request.