The automatic reconstruction of three-dimensional models from acquired data, whether images or 3D point clouds, has been one of the central themes of computer graphics and computer vision for several decades.
In this field, CRS4 has introduced various cutting-edge visual computing technologies, using both geometric reasoning and deep learning approaches based on artificial neural networks. In 2016, in particular, one of the first methods capable of effectively reconstructing multi-room interiors from point clouds was proposed. Later, we focused on the even more challenging task of reconstructing scenes from panoramic images, an area in which in recent years we have developed several cutting-edge solutions.
Relevant publications on this subject are the following:
- Claudio Mura, Oliver Mattausch, Alberto Jaspe Villanueva, Enrico Gobbetti, and Renato Pajarola. Automatic Room Detection and Reconstruction in Cluttered Indoor Environments with Complex Room Layouts. Computers & Graphics, 44: 20-32, November 2014. https://doi.org/10.1016/j.cag.2014.07.005.
- Giovanni Pintore, Fabio Ganovelli, Ruggero Pintus, Roberto Scopigno, and Enrico Gobbetti. 3D floor plan recovery from overlapping spherical images. Computational Visual Media, 4(4): 367-383, December 2018. https://doi.org/10.1007/s41095-018-0125-9.
- Giovanni Pintore, Ruggero Pintus, Fabio Ganovelli, Roberto Scopigno, and Enrico Gobbetti. Recovering 3D existing-conditions of indoor structures from spherical images. Computers & Graphics, 77: 16-29, December 2018. https://doi.org/10.1016/j.cag.2018.09.013.
- Giovanni Pintore, Fabio Ganovelli, Alberto Jaspe Villanueva, and Enrico Gobbetti. Automatic modeling of cluttered multi-room floor plans from panoramic images. Computers Graphics Forum, 38(7): 347-358, 2019. https://doi.org/10.1111/cgf.13842
- Giovanni Pintore, Marco Agus, and Enrico Gobbetti. AtlantaNet: Inferring the 3D Indoor Layout from a Single 360 Image beyond the Manhattan World Assumption. In Proc. ECCV, August 2020.