From Coarse to Fine: Efficient Training for Audio Spectrogram Transformers.

Jiu Feng Mehmet Hamza Erol Joon Son Chung Arda Senocak

Published in: CoRR (2024)

Keyphrases

coarse to fine
multiresolution
multiscale
hierarchical segmentation
image registration
hierarchical representation
object detection
optical flow estimation
matching scheme
scan line
neural network
machine learning
image segmentation
training set
wavelet transform
deformable contour