Improving Audiovisual Active Speaker Detection in Egocentric Recordings with the Data-Efficient Image Transformer.

Jason Clarke Yoshihiko Gotoh Stefan Goetze

Published in: ASRU (2023)

Keyphrases

image data
data sets
data sources
high quality
image classification
input image
video recordings
audio visual
image analysis
digital images
satellite images
image regions
low level
training data
image features
data points
image quality
multi modal
image representation
single image
image retrieval
visual data
data structure