In this paper, we propose an image adjustment method for multi-exposure images based on convolutional neural networks (CNNs). We call image regions without information due to saturation and object moving in multi-exposure images lacking areas in this paper. Lacking areas cause the ghosting artifact in fused images from sets of multi-exposure images by conventional fusion methods, which tackle the artifact. To avoid this problem, the proposed method estimates the information of lacking areas via adaptive inpainting. The proposed CNN consists of three networks, warp and refinement, detection, and inpainting networks. The second and third networks detect lacking areas and estimate their pixel values, respectively. In the experiments, it is observed that a simple fusion method with the proposed method outperforms state-of-the-art fusion methods in the peak signal-to-noise ratio. Moreover, the proposed method is applied for various fusion methods as pre-processing, and results show obviously reducing artifacts.
Isana FUNAHASHI
Univ. of Electro-Commun.
Taichi YOSHIDA
Univ. of Electro-Commun.
Xi ZHANG
Univ. of Electro-Commun.
Masahiro IWAHASHI
Nagaoka Univ. of Tech.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Isana FUNAHASHI, Taichi YOSHIDA, Xi ZHANG, Masahiro IWAHASHI, "Image Adjustment for Multi-Exposure Images Based on Convolutional Neural Networks" in IEICE TRANSACTIONS on Information,
vol. E105-D, no. 1, pp. 123-133, January 2022, doi: 10.1587/transinf.2021EDP7087.
Abstract: In this paper, we propose an image adjustment method for multi-exposure images based on convolutional neural networks (CNNs). We call image regions without information due to saturation and object moving in multi-exposure images lacking areas in this paper. Lacking areas cause the ghosting artifact in fused images from sets of multi-exposure images by conventional fusion methods, which tackle the artifact. To avoid this problem, the proposed method estimates the information of lacking areas via adaptive inpainting. The proposed CNN consists of three networks, warp and refinement, detection, and inpainting networks. The second and third networks detect lacking areas and estimate their pixel values, respectively. In the experiments, it is observed that a simple fusion method with the proposed method outperforms state-of-the-art fusion methods in the peak signal-to-noise ratio. Moreover, the proposed method is applied for various fusion methods as pre-processing, and results show obviously reducing artifacts.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDP7087/_p
Copy
@ARTICLE{e105-d_1_123,
author={Isana FUNAHASHI, Taichi YOSHIDA, Xi ZHANG, Masahiro IWAHASHI, },
journal={IEICE TRANSACTIONS on Information},
title={Image Adjustment for Multi-Exposure Images Based on Convolutional Neural Networks},
year={2022},
volume={E105-D},
number={1},
pages={123-133},
abstract={In this paper, we propose an image adjustment method for multi-exposure images based on convolutional neural networks (CNNs). We call image regions without information due to saturation and object moving in multi-exposure images lacking areas in this paper. Lacking areas cause the ghosting artifact in fused images from sets of multi-exposure images by conventional fusion methods, which tackle the artifact. To avoid this problem, the proposed method estimates the information of lacking areas via adaptive inpainting. The proposed CNN consists of three networks, warp and refinement, detection, and inpainting networks. The second and third networks detect lacking areas and estimate their pixel values, respectively. In the experiments, it is observed that a simple fusion method with the proposed method outperforms state-of-the-art fusion methods in the peak signal-to-noise ratio. Moreover, the proposed method is applied for various fusion methods as pre-processing, and results show obviously reducing artifacts.},
keywords={},
doi={10.1587/transinf.2021EDP7087},
ISSN={1745-1361},
month={January},}
Copy
TY - JOUR
TI - Image Adjustment for Multi-Exposure Images Based on Convolutional Neural Networks
T2 - IEICE TRANSACTIONS on Information
SP - 123
EP - 133
AU - Isana FUNAHASHI
AU - Taichi YOSHIDA
AU - Xi ZHANG
AU - Masahiro IWAHASHI
PY - 2022
DO - 10.1587/transinf.2021EDP7087
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E105-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2022
AB - In this paper, we propose an image adjustment method for multi-exposure images based on convolutional neural networks (CNNs). We call image regions without information due to saturation and object moving in multi-exposure images lacking areas in this paper. Lacking areas cause the ghosting artifact in fused images from sets of multi-exposure images by conventional fusion methods, which tackle the artifact. To avoid this problem, the proposed method estimates the information of lacking areas via adaptive inpainting. The proposed CNN consists of three networks, warp and refinement, detection, and inpainting networks. The second and third networks detect lacking areas and estimate their pixel values, respectively. In the experiments, it is observed that a simple fusion method with the proposed method outperforms state-of-the-art fusion methods in the peak signal-to-noise ratio. Moreover, the proposed method is applied for various fusion methods as pre-processing, and results show obviously reducing artifacts.
ER -