GTF: Omnidirectional EPI Transformer for Light Field Super-Resolution

TL;DR

This paper introduces GTF, a novel omnidirectional EPI Transformer for light field super-resolution that models all directional EPIs, including diagonal ones, to improve reconstruction quality.

Contribution

GTF is the first unified framework explicitly modeling horizontal, vertical, and diagonal EPIs, enhancing light field super-resolution performance and efficiency.

Findings

GTF achieves 32.78 dB on five LF SR benchmarks.

GTF-Tiny attains 32.57 dB with only 0.915M parameters.

Submissions ranked 3rd and 4th in NTIRE 2026 LF SR challenge.

Abstract

Light field (LF) image super-resolution benefits from Epipolar Plane Images (EPIs), whose line slopes explicitly encode disparity. However, existing Transformer-based LF SR methods mainly attend to horizontal and vertical EPIs, leaving diagonal epipolar geometry underexplored. We present GTF, an omnidirectional EPI Transformer that explicitly models horizontal, vertical, 45-degree, and 135-degree EPIs within a unified reconstruction framework. GTF combines directional EPI processing, MacPI-based prior injection, adaptive directional fusion, and a topology-preserving feed-forward network to better exploit LF geometry. For the NTIRE 2026 fidelity tracks, we use GTF as the main model, while a lightweight GTF-Tiny variant targets the efficiency track. On five standard LF SR benchmarks covering both real-captured and synthetic scenes, GTF reaches 32.78 dB without inference-time enhancement, and stronger inference settings with EPSW and test-time augmentation further improve performance. Under the NTIRE 2026 efficiency constraint, GTF-Tiny attains 32.57 dB with only 0.915M parameters and 19.81 GFLOPs. In the NTIRE 2026 Light Field Image Super-Resolution Challenge, our submissions rank 3rd on Track 1 and Track 3 and 4th on Track 2. Architecture-evolution, channel-width, and inference analyses further support the effectiveness of diagonal EPI modeling, directional fusion, and the lightweight design.