Joint Feature Aggregation for Stereo Image Super-resolution

Zekun Ai,Xiaotong Luo,Yanyun Qu

DOI: https://doi.org/10.1109/ICME55011.2023.00089

2023-01-01

Abstract:Stereo image super-resolution (Stereo SR) has been a newly rising and challenging problem with the popular application of dual cameras, which can be used to promote the SR performance by adding auxiliary information from another viewpoint. Most of the existing excellent works have concentrated on leveraging the intrinsic feature correlation of two view images via exploring the non-local attention mechanism. However, they only perform interaction once for feature registration and fusion accompanied by the complex view transition constraint, which cannot fully take advantage of the information in the stereo image pairs. In this paper, we propose a joint feature aggregation network for Stereo SR to calibrate single-view features and integrate cross-view knowledge effectively. Specifically, we introduce a self-calibrated feature extractor to excavate multi-scale and multi-direction features within the single-view image. What's more, we design an adaptive fusion module with the cross-view attention mechanism, which is utilized to mine and fuse the long-range dependencies between the stereo image pairs so as to get rid of the inflexible cycle constraints. Extensive experimental results demonstrate that our proposal successfully achieves superior performance against the state-of-the-art methods on four datasets.