Reproduce “Proactive tactile exploration for object-agnostic shape reconstruction from minimal visual priors” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Paris Oikonomou, George Retsinas, Petros Maragos, Costas S. Tzafestas

An iterative methodology reconstructs 3D object shapes from minimal visual priors by combining global template fitting with local mesh deformation and a proactive tactile exploration strategy. This approach improves reconstruction accuracy and reduces contact failures by balancing shape uncertainty reduction with contact reliability.

Key insights

  • Combining a coarse global shape estimation with local mesh deformation effectively allows for robust 3D reconstruction even with minimal initial visual data.
  • Explicitly incorporating the risk of contact failure into the tactile exploration strategy, alongside shape uncertainty, significantly enhances exploration efficiency and robustness in physical interaction.
  • Dynamically calculating tactile uncertainty based on contact force and torque provides more precise guidance for exploration, particularly in distinguishing reliable contacts from those at the sensor's edge.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.