# ScrewNet: Category-Independent Articulation Model Estimation From Depth Images Using Screw Theory Source: [ScrewNet: Category-Independent Articulation Model Estimation From Depth Images Using Screw Theory]() Collection index: [llms.txt]() Data updated: 2025-10-25T23:23:37.152Z Authors: Ajinkya Jain, Rudolf Lioutikov, Caleb Chuck, Scott Niekum Published: 2020-08-24T15:41:23.000Z ## Abstract Robots in human environments will need to interact with a wide variety of articulated objects such as cabinets, drawers, and dishwashers while assisting humans in performing day-to-day tasks. Existing methods either require objects to be textured or need to know the articulation model category a priori for estimating the model parameters for an articulated object. We propose ScrewNet, a novel approach that estimates an object's articulation model directly from depth images without requiring a priori knowledge of the articulation model category. ScrewNet uses screw theory to unify the representation of different articulation types and perform category-independent articulation model estimation. We evaluate our approach on two benchmarking datasets and compare its performance with a current state-of-the-art method. Results demonstrate that ScrewNet can successfully estimate the articulation models and their parameters for novel objects across articulation model categories with better on average accuracy than the prior state-of-the-art method. Project webpage: https://pearl-utexas.github.io/ScrewNet/ ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Geometry Reconstruction on Synthetic Dataset](): Geometry Reconstruction; Synthetic dataset; Whole Chamfer Distance: 0.54. - [Articulation Estimation on Synthetic Dataset](): Articulation Estimation; Synthetic dataset; Prismatic Angle Error: 0.69. - [Articulated Object Joint Estimation on Real world dataset Unseen categories 1.0](): Articulated Object Joint Estimation; Real world dataset Unseen categories 1.0; Axis Orientation Error (°): 28.22. - [Geometry Reconstruction on Shape2Motion](): Geometry Reconstruction; Shape2Motion; Whole Chamfer Distance: 0.9. - [Articulation Estimation on Shape2Motion](): Articulation Estimation; Shape2Motion; Prismatic Angle Error: 1.36. - [Articulated Object Joint Estimation on Real world dataset Seen categories 1.0 (seen\_objects)](): Articulated Object Joint Estimation; Real world dataset Seen categories 1.0 (seen\_objects); Axis Orientation Error (°): 28.42. - [Articulated Object Joint Estimation on PartNet-Mobility Simulated Dataset Unseen Categories v0 (test)](): Articulated Object Joint Estimation; PartNet-Mobility Simulated Dataset Unseen Categories v0 (test); Orientation Error (deg): 25.72. - [Articulated Object Joint Estimation on PartNet-Mobility Simulated Dataset Seen Categories v0 (test)](): Articulated Object Joint Estimation; PartNet-Mobility Simulated Dataset Seen Categories v0 (test); Orientation Error (°): 20.74. ## Original sources and implementation [Original paper]() [PDF]()