SHREC 2022: Fitting and recognition of simple geometric primitives on point clouds

The aim of this SHREC track is to evaluate the quality of automatic algorithms for fitting and recognising geometric primitives in point clouds. The goal is to identify, for each point cloud, its primitive type and some geometric descriptors. As simple primitives we mean the classical surface primitives derived from constructive solid geometry, i.e., planes, spheres, cylinders, cones and tori. We admit that some point clouds might be unfitted by one of these simple primitives.

The track is organized by IMATI-CNR. The final report of this SHREC track will be submitted as a joint contribution to the international journal Computers & Graphics and will follow a two-stage review process. The paper will be authored by the track coordinators and all participants who submitted their results.

Fitting and recognition are of paramount importance in multiple application domains, such as, in reverse engineering and data compression. In most real-world scenarios, acquisition methods tend to produce data affected by noise and potentially other imperfections (e.g., non-uniform or low sampling density, misalignment and missing data). The literature on methods for fitting and recognition of simple primitives is quite consolidated, and it is therefore worth analyzing and comparing their outcome in terms of different evaluation measures. For this purpose we create a synthetic dataset containing segments perturbed with different kinds of point cloud artifacts.

Description of the track
A dataset of 3D segments represented as point clouds, divided in training set and test set, will be provided. For each segment in the training set we will make available its primitive type and an analytical representation of it. For each point cloud in the test set, the goal of the track is twofold: recognise the primitive type from the data and provide an analytical representation of the primitive recognised.

The relevance or non-relevance of the simple primitives will be evaluated on the basis of an existing ground truth. We will generate the dataset sampling parametric equations and subsequently applying different kinds of point cloud perturbations (e.g., different types of undersampling, noise and outliers).

Performance evaluation

The performance of the algorithms will be evaluated using both classification and approximation measures (to quantify the goodness of fit and the quality of the inferred descriptors). This analysis will be specified according to the type of point cloud artifacts, in order to provide a thorough evaluation of the robustness of the algorithms.

Registration and other procedures
Each participant is requested to register to the track by sending an email to Chiara Romanengo (email: and Andrea Raffo (email: with the subject "SHREC 2022: Fitting and recognition of simple geometric primitives on point clouds". Then, an answer will be sent to each participant with further instructions.

We plan to submit the benchmark code and data to the replicability stamp

Further Information

Important dates