Be Everywhere - Hear Everything (BEE): Audio Scene Reconstruction by Sparse Audio-Visual Samples.
Mingfei ChenKun SuEli ShlizermanPublished in: ICCV (2023)
Keyphrases
- audio visual
- scene reconstruction
- multi view
- multi modal
- bundle adjustment
- visual data
- visual information
- image based rendering
- problems in computer vision
- scene structure
- camera motion
- audio features
- high dimensional
- camera parameters
- audio visual speech recognition
- multi stream
- camera calibration
- multimedia
- fundamental matrix
- data sets
- image correspondences
- view synthesis
- structure from motion
- uncalibrated cameras
- multimodal fusion
- multi camera
- d scene
- pairwise
- training set
- computer vision