An artificial intelligence framework is described that incorporates a number of neural networks and a number of transformers for converting a two-dimensional image into three-dimensional semantic information. Neural networks convert one or more images into a set of image feature maps, depth information associated with the one or more images, and query proposals based on the depth information. A first transformer implements a cross-attention mechanism to process the set of image feature maps in accordance with the query proposals. The output of the first transformer is combined with a mask token to generate initial voxel features of the scene. A second transformer implements a self-attention mechanism to convert the initial voxel features into refined voxel features, which are up-sampled and processed by a lightweight neural network to generate the three-dimensional semantic information, which may be used by, e.g., an autonomous vehicle for various advanced driver assistance system (ADAS) functions.
SPARSE VOXEL TRANSFORMER FOR CAMERA-BASED 3D SEMANTIC SCENE COMPLETION
2024-03-14
Patent
Electronic Resource
English
IPC: | G06T Bilddatenverarbeitung oder Bilddatenerzeugung allgemein , IMAGE DATA PROCESSING OR GENERATION, IN GENERAL / B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion / G06V |
Instance-Aware Monocular 3D Semantic Scene Completion
IEEE | 2024
|Photorealistic Scene Reconstruction by Voxel Coloring
British Library Online Contents | 1999
|Monocular localization within sparse voxel maps
IEEE | 2017
|