EiffHDR: An Efficient Network for Multi-Exposure High Dynamic Range Imaging

While recent progress in Multi-exposure HDR imaging is promising, the growing complexity of state-of-the-art (SOTA) methods poses challenges for their analysis and comparison. In this paper, we analyze the motivations and approaches behind previous SOTA works and introduce EiffHDR, an efficient Multi-exposure HDR imaging technique. In contrast to prior methods employing multiple branches spatial attention mechanisms, EiffHDR adopts a streamlined gating mechanism for information flow control at both spatial and channel levels, enabling implicit alignment. Subsequently, we process these features through proposed Efficient Merging Network, facilitating long-range correlations and multi-scale information perception, ultimately producing high-quality HDR images. Our experiments demonstrate that EiffHDR not only achieves outstanding performance but also significantly reduces computational complexity, making it a valuable contribution to the field.

1. INTRUDUCTION

In response to common issues in digital photography, where Low Dynamic Range (LDR) images often suffer from overexposure or underexposure, a typical practice involves capturing multiple LDR images with varying exposures and then merging them into a High Dynamic Range (HDR) image. However, this approach faces two significant challenges as discussed in the literature [1]. First, misalignment among LDR images results in artifacts and ghosting in the final output. Second, inadequate image information arises from saturated regions in LDR inputs.