Assessing Learned Models for Phase-only Hologram Compression

TL;DR

This paper evaluates various learned models for phase-only hologram compression, highlighting the effectiveness of INRs like SIREN and the limitations of pretrained VAEs such as TAESD in this specific task.

Contribution

It provides a comparative analysis of INR and VAE models for hologram compression, revealing the strengths of SIREN and the need for task-specific VAE adaptations.

Findings

SIREN achieves 40% compression with high quality (PSNR = 34.54 dB)

Pretrained image VAEs like TAESD struggle with hologram compression

INRs outperform pretrained VAEs in this application

Abstract

We evaluate the performance of four common learned models utilizing INR and VAE structures for compressing phase-only holograms in holographic displays. The evaluated models include a vanilla MLP, SIREN, and FilmSIREN, with TAESD as the representative VAE model. Our experiments reveal that a pretrained image VAE, TAESD, with 2.2M parameters struggles with phase-only hologram compression, revealing the need for task-specific adaptations. Among the INRs, SIREN with 4.9k parameters achieves %40 compression with high quality in the reconstructed 3D images (PSNR = 34.54 dB). These results emphasize the effectiveness of INRs and identify the limitations of pretrained image compression VAEs for hologram compression task.