CommIN: semantic image communications as an inverse problem with INN-guided diffusion models
File(s)
Accepted version
Author(s)
Chen, Jiakang
You, Di
Gündüz, Deniz
Dragotti, Pier Luigi
Type
Conference Paper
Abstract
Joint source-channel coding schemes based on deep neural networks (DeepJSCC) have recently achieved remarkable performance for wireless image transmission. However, these methods usually focus only on the distortion of the reconstructed signal at the receiver side with respect to the source at the transmitter side, rather than the perceptual quality of the reconstruction which carries more semantic information. As a result, severe perceptual distortion can be introduced under extreme conditions such as low bandwidth and low signal-to-noise ratio. In this work, we propose CommIN, which views the recovery of high-quality source images from degraded reconstructions as an inverse problem. To address this, CommIN combines Invertible Neural Networks (INN) with diffusion models, aiming for superior perceptual quality. Through experiments, we show that our CommIN significantly improves the perceptual quality compared to DeepJSCC under extreme conditions and outperforms other inverse problem approaches used in DeepJSCC.
Date Issued
2024-03-18
Date Acceptance
2024-04-01
Citation
ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024, pp.6675-6679
ISBN
979-8-3503-4485-1
ISSN
1520-6149
Publisher
IEEE
Start Page
6675
End Page
6679
Journal / Book Title
ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Copyright Statement
© 2024 IEEE. This is the author’s accepted manuscript made available under a CC-BY licence in accordance with Imperial’s Research Publications Open Access policy (www.imperial.ac.uk/oa-policy)
License URL
Source
ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publication Status
Published
Start Date
2024-04-14
Finish Date
2024-04-19
Coverage Spatial
Seoul, Korea
Date Publish Online
2024-03-18