OSLO-IC: On-the-Sphere Learned Omnidirectional Image Compression with Attention Modules and Spatial Context

Wawerek-López P, Mahmoudian Bidgoli N, Fossard P, Kaup A, Maugey T (2025)


Publication Language: English

Publication Type: Conference contribution

Publication year: 2025

Publisher: IEEE

City/Town: New York City

Pages Range: 1-5

Event location: Hyderabad IN

URI: https://ieeexplore.ieee.org/document/10889131

DOI: 10.1109/ICASSP49660.2025.10889131

Open Access Link: https://arxiv.org/abs/2503.13119

Abstract

Developing effective 360-degree (spherical) image compression techniques is crucial for technologies like virtual reality and automated driving. This paper advances the state-of-the-art in on-the-sphere learning (OSLO) for omnidirectional image compression framework by proposing spherical attention modules, residual blocks, and a spatial autoregressive context model. These improvements achieve a 23.1% bit rate reduction in terms of WS-PSNR BD rate. Additionally, we introduce a spherical transposed convolution operator for upsampling, which reduces trainable parameters by a factor of four compared to the pixel shuffling used in the OSLO framework, while maintaining similar compression performance. Therefore, in total, our proposed method offers significant rate savings with a smaller architecture and can be applied to any spherical convolutional application.

Authors with CRIS profile

Involved external institutions

How to cite

APA:

Wawerek-López, P., Mahmoudian Bidgoli, N., Fossard, P., Kaup, A., & Maugey, T. (2025). OSLO-IC: On-the-Sphere Learned Omnidirectional Image Compression with Attention Modules and Spatial Context. In Proceedings of the ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 1-5). Hyderabad, IN.

MLA:

Wawerek-López, Paul, et al. "OSLO-IC: On-the-Sphere Learned Omnidirectional Image Compression with Attention Modules and Spatial Context." Proceedings of the ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad 2025. 1-5.

BibTeX: Download