Kathirvel, RP and Agrawal, S and Radhakrishnan, VB (2021) Self-Gated Memory Recurrent Network for Efficient Scalable HDR Deghosting. In: IEEE Transactions on Computational Imaging, 7 . pp. 1228-1239.
|
PDF
IEEE_tra_com_ima_ope_7_1228-1239_2021.pdf - Submitted Version Download (5MB) | Preview |
Abstract
We propose a novel recurrent network-based HDR deghosting method for fusing arbitrary length dynamic sequences. The proposed method uses convolutional and recurrent architectures to generate visually pleasing, ghosting-free HDR images. We introduce a new recurrent cell architecture, namely Self-Gated Memory (SGM) cell, that outperforms the standard LSTM cell while containing fewer parameters and having faster running times. In the SGM cell, the information flow through a gate is controlled by multiplying the gate's output by a function of itself. Additionally, we use two SGM cells in a bidirectional setting to improve output quality. The proposed approach achieves state-of-the-art performance compared to existing HDR deghosting methods quantitatively across three publicly available datasets while simultaneously achieving scalability to fuse variable length input sequence without necessitating re-training. Through extensive ablations, we demonstrate the importance of individual components in our proposed approach.
| Item Type: | Journal Article |
|---|---|
| Publication: | IEEE Transactions on Computational Imaging |
| Publisher: | Institute of Electrical and Electronics Engineers Inc. |
| Additional Information: | The copyright for this article belongs to the Authors. |
| Keywords: | computational photography; convolutional neural networks; deghosting; exposure fusion; High dynamic range image fusion |
| Department/Centre: | Division of Interdisciplinary Sciences > Computational and Data Sciences |
| Date Deposited: | 06 Jun 2023 10:09 |
| Last Modified: | 06 Jun 2023 10:09 |
| URI: | https://eprints.iisc.ac.in/id/eprint/81817 |
Actions (login required)
 |
View Item |
