Work

Indoor Semantic Segmentation using depth information

Year: 2013

Type: conference-paper

Abstract: This work addresses multi-class segmentation of indoor scenes with RGB-D inputs. While this area of research has gained much attention recently, most works still rely on hand-crafted features. In contrast, we apply a multiscale convolutional network to learn features directly from the images and the depth information. We obtain state-of-the-art on the NYU-v2 depth dataset with an accuracy of 64.5%. We illustrate the labeling of indoor scenes in videos sequences that could be processed in real-time using appropriate hardware such as an FPGA.

Source: HAL (Le Centre pour la Communication Scientifique Directe)

DOI: 10.48550/arxiv.1301.3572

Authors: Couprie, Camille, Farabet, Clément, Najman, Laurent, LeCun, Yann

Institutions: IFP Énergies nouvelles, Courant Institute of Mathematical Sciences, Laboratoire d'Informatique Gaspard-Monge, ESIEE Paris, Pôle de recherche et d'enseignement supérieur Université Paris-Est, New York University

Language: English

Topic: Advanced Neural Network Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W2963038646

FWCI: 19.58

Cites:

Cited by: 393

Related to: 10