Light field (LF) imaging captures spatial and angular information, offering a 4D scene representation enabling enhanced visual understanding. However, high dimensionality and redundancy across spatial and angular domains present major challenges for compression, particularly where storage, transmission bandwidth, or processing latency are constrained. 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 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 latent feature specialization 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 relying 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 our method achieves strong compression performance, yielding an average BD-PSNR gain of 2.91 dB over HEVC and high compression ratios (e.g., 200:1). Additionally, our design enables fast inference, with a total end-to-end time over 19x faster than the JPEG Pleno standard, 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, effective solution for practical light field image compression.

Memristor-based neural networks promise low-power, high-density neuromorphic hardware, yet most studies remain proof-of-concept with small inputs and qualitative reporting. We present a comprehensive investigation of on-chip learning in a scalable memristor network, modeling realistic non-idealities including device-to-device and cycle-to-cycle variability, conductance programming errors, and input noise. Using a 30-memristor architecture, we systematically scale inputs from 3×3 to 11x11 and evaluate classification robustness under noisy conditions. Results show that moderate training noise improves tolerance, enabling up to 99% accuracy with 20% noisy tests, whereas aggressive weight quantization at higher resolutions can degrade accuracy. We quantify average training time and energy across device types; chromium devices are most efficient, achieving 2.4 s average training time and 18.9mJ training energy for 3×3 tasks. We also report energy per classification versus input size, establishing concrete benchmarks often absent from prior work. A Big-O analysis indicates sub-linear growth of training time (n0.26), training energy (n0.28), and inference cost with input dimension, supporting suitability for edge deployment. Unlike earlier demonstrations that emphasized qualitative recognition, this study provides quantitative trade-offs among accuracy, noise robustness, energy, and training time, bridging device-level feasibility and practical memristor neural network accelerators. These results guide design choices for deployable edge systems.

Light fields (LF) technology enables the capture and reproduction of a 3D scene accurately, which enhances visual experience in various applications. The sheer volume of multiplicity of the captured views creates logistical problems in both storage and data transmission, which makes LF compression crucial. Even though many LF compression techniques have been proposed and evaluated in recent years, the leading-edge learning-based approaches have not been subject to the same level of scrutiny. This paper presents a subjective quality assessment study on four different LF compression methods, including two learning-based LF compression methods, which have not been studied before from a subjective quality point of view. For this purpose, subjective opinion scores were collected from viewers in two different universities for a cross-lab study. The results indicate that the learning-based compression methods have different behavior in their rate-distortion curves, and that there is room for improvement for learning-based methods. A qualitative analysis also shows that their artifact structures are different from conventional ones. The results highlight the need for a perceptual objective quality metric that takes different types of artifacts into account. The obtained subjective quality database (MiX-LFQDB) is made public to support further research in this area: https://doi.org/10.5281/zenodo.16778670

Recent advancements in learned light field (LF) image compression highlight the advantages of modeling spatial and angular redundancies using deep generative models. Among these, Variational Autoencoders (VAEs) have shown strong potential in learning compact latent representations of LF data. However, many existing approaches rely on simple uniform quantization and basic entropy coding, which limits compression efficiency in practical applications. This paper introduces DUALF-C, a lightweight and retraining-free compression pipeline that augments pretrained VAE-based LF compression models using structured post-encoding transformations. The proposed framework integrates bitplane slicing, latent channel reordering, non-uniform quantization, and patch-based vector quantization to improve bitrate efficiency while preserving reconstruction quality. Experimental evaluations demonstrate that DUALF-C significantly reduces bit-per-pixel (BPP) without degrading image quality, making it a practical solution for bandwidth-constrained immersive imaging systems.

Immersive imaging technologies provide an enhanced user experience for visual applications and are getting ready for commonplace use by the industry and the general populace. In particular, light field is a promising technology that enables the capture and reproduction of real light rays from the scene, which can provide a backbone for immersive telecommunication and telepresence applications. Nevertheless, there are still many challenges in transmitting and reproducing light field data. This paper proposes a spherical light field dataset that can be used as a foundation for developing telepresence applications. The Spherical Light Field Database (SLFDB) consists of a light field of 60 views captured with an omnidirectional camera in 20 scenes. To show the usefulness of the proposed database, we provide two use cases: compression and viewpoint estimation. The initial results validate that the publicly available SLFDB will benefit the scientific community.