Work

DUALF-D: Disentangled dual-hyperprior approach for light field image compression

Year: 2025

Type: article

Abstract: Light field (LF) imaging captures both spatial and angular information, offering a rich 4D scene representation that enables enhanced visual understanding. However, the high dimensionality and redundancy across spatial and angular domains present major challenges for compression, particularly in scenarios where storage, transmission bandwidth, or processing latency are constrained. In this paper, we present a novel Variational Autoencoder (VAE)-based framework that explicitly disentangles spatial and angular features using two parallel latent branches. Each branch is coupled with an independent hyperprior model, allowing for more precise distribution estimation for entropy coding and finer rate-distortion control. This dual-hyperprior structure enables the network to adaptively compress spatial and angular information based on their unique statistical characteristics, improving coding efficiency. To further enhance the specialization of latent features and promote disentanglement, we introduce a mutual information-based regularization term that minimizes redundancy between the two branches while preserving feature diversity. Unlike prior methods that rely on covariance-based penalties prone to collapse, our information-theoretic regularizer provides more stable and interpretable latent separation. Experimental results on publicly available LF datasets demonstrate that our method achieves superior compression performance compared to state-of-the-art methods, while maintaining competitive reconstruction quality. In addition, our design enables fast inference time, making it well-suited for real-time and bandwidth-sensitive applications. By jointly leveraging disentangled representation learning, dual-hyperprior modeling, and information-theoretic regularization, our approach offers a scalable and effective solution for practical light field image compression.

Source: Signal Processing Image Communication

DOI: 10.1016/j.image.2025.117436

Authors: Soheib Takhtardeshir, Roger Olsson, Christine Guillemot, Mårten Sjöström

Institutions: Institut national de recherche en sciences et technologies du numérique, Centre Inria de l'Université de Rennes, Mid Sweden University

Language: English

Topic: Advanced Data Compression Techniques

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $2,470