IRIS-VIS

A New Dataset for Visibility Estimation in an Industrial Environment

Flavien Armangeon1,2, Thibaud Ehret1, Enric Meinhardt-Llopis1 Rafael Grompone Von Gioi1, Guillaume Thibault2, Marc Petit2, Gabriele Facciolo1,
1Université Paris-Saclay, CNRS, ENS Paris-Saclay, Centre Borelli, France 2EDF R&D (Electricité de France), France
Paper IRIS Dataset Visibility Estimation Data Code
MY ALT TEXT

Abstract

Point cloud visibility estimation is fundamental as it is useful for many computer vision applications including surface reconstruction, 3D segmentation from paired images and point densification. Previous works showed outstanding results on simple object and outdoor datasets. However, unlike the previously studied scenes, the most challenging environments are those providing a high amount of object points in the same direction, typically in complex indoor scenes. In this kind of environments, due to the lack of real data ground truth, quantitative analysis are either missing or based on simulated data. In this work, we present IRIS-VIS (Industrial Room In Saclay - VISibility), a new dataset for point visibility estimation in an indoor environment. It is a high complexity scene due to the large variety in the shape, size and orientation of the objects. To our knowledge, this is the first dataset on real indoor data providing a dense LiDAR station-based point cloud along with a well-fitted CAD model. The latter is useful to compute automatically, quickly and accurately the visibility from any given viewpoint, enabling evaluations under infinite conditions. We propose new metrics for the visibility estimation task and evaluate state-of-the-art methods in both sparse and dense conditions with the proposed dataset.

Conference Poster

Visibility Estimation Data

IRIS provides a dense LiDAR point cloud and a CAD model reconstructed close to the points. IRIS-VIS adds the point visibility ground truth and the complex visibility areas that be computed from any viewpoint thanks to the the model using our code. Qualitative clouds, corrected mesh and complex areas examples are also given. More details can be found on the IRIS page and download page.

Ground truth constuction

The mesh given by the CAD model (violet) is paired with the cloud. As the model is reconstructed close to the points, we can use the paired mesh (orange) as the reference for the visibility by simple raycasting (see paper for more details).

Complex visibility metrics

We designed new metrics for the visibility estimation task to focus on the areas where the variability of the visible points is high.

BibTeX


      @InProceedings{Armangeon_2025_WACV,
        author    = {Armangeon, Flavien and Ehret, Thibaud and Meinhardt-Llopis, Enric and von Gioi, Rafael Grompone and Thibault, Guillaume and Petit, Marc and Facciolo, Gabriele},
        title     = {IRIS-VIS: A New Dataset for Visibility Estimation in an Industrial Environment},
        booktitle = {Proceedings of the Winter Conference on Applications of Computer Vision (WACV)},
        month     = {February},
        year      = {2025},
        pages     = {7235-7243}
      }