Synthetic benchmark enables tests of deformable 3D shape reconstruction
Synthetic Deformable Tubular Benchmark: A mechanically annotated dataset for evaluating deformable 3D reconstruction
Computer Vision and Pattern Recognition
Summary
Reconstructing 3D shapes that bend and move is hard to check without detailed known references. The authors created a synthetic dataset with images and full mechanical annotations of deforming tubular shapes seen from different camera angles. This dataset lets users test and compare methods that track surface motion, shape, and material points over time. It helps understand how camera movement and physical deformation impact reconstruction performance by offering detailed ground truth data. The authors also provide tools to evaluate annotations and extend the dataset.
What this means in practice
- •For computer vision developers: Benchmark algorithms that reconstruct deforming 3D shapes from images with full known ground truth on shape and motion.
- •For graphics software engineers: Test and improve physically based animation and tracking methods by using a detailed, annotated synthetic deformable tubular dataset.
Authors
Íñigo Cirauqui-Viloria
Abstract
Evaluating deformable 3D reconstruction requires ground truth linking camera motion, surface deformation, material correspondence and image appearance. We introduce the Synthetic Deformable Tubular Benchmark, generated with LumenMorph, an open-source toolkit for creating, annotating and evaluating tubular scenes. The release contains 5,580 RGB images. Its baseline crosses three surface-motion conditions with moving and stationary cameras in six scene families; matched appearance variants cover only moving-camera finite-element actions in four families. Documented public and evaluation interfaces expose camera poses, geometry, depth, deformation fields, persistent material correspondences and discrete finite-element records. Persistent triangle identities and material weights let users track the same surface location across frames and query its three-dimensional trajectory. Separately implemented geometric and mechanical checks assess annotation consistency with the generating models. Structure-from-motion, optical-flow and supplied-geometry displacement-interpolation studies demonstrate how the resource distinguishes registration coverage from geometric accuracy, exposes camera- and surface-motion effects, and tests the benefits and mismatch of mechanical priors. Public readers, evaluation contracts and the configurable generator support reuse and extension. Data integrity, fixed-environment repeatability and retained-prediction score replay have distinct verification scopes. The benchmark provides a controlled synthetic reference without claiming anatomical or clinical realism.