A novel deep-learning based depth estimation method for light field images is introduced. The proposed method employs a novel neural network design to estimate the disparity of each pixel based on block patches extracted from Epipolar Plane Images. The network output is further refined based on filtering and denoising algorithms. Experimental results demonstrate an average improvement of 34.35% in RMSE and 49.44% in MSE over machine learning based state-of-the-art methods.
Schiopu, I & Munteanu, A 2019, 'Deep-Learning based Depth Estimation for Light Field Images', Electronics Letters, vol. 55, no. 20, el.2019.2073, pp. 1086–1088. https://doi.org/10.1049/el.2019.1757
Schiopu, I., & Munteanu, A. (2019). Deep-Learning based Depth Estimation for Light Field Images. Electronics Letters, 55(20), 1086–1088. Article el.2019.2073. https://doi.org/10.1049/el.2019.1757
@article{c2939945a2014d5a8d56a50c1810f300,
title = "Deep-Learning based Depth Estimation for Light Field Images",
abstract = "A novel deep-learning based depth estimation method for light field images is introduced. The proposed method employs a novel neural network design to estimate the disparity of each pixel based on block patches extracted from Epipolar Plane Images. The network output is further refined based on filtering and denoising algorithms. Experimental results demonstrate an average improvement of 34.35% in RMSE and 49.44% in MSE over machine learning based state-of-the-art methods.",
keywords = "Deep-learning, Depth estimation, Light field images",
author = "Ionut Schiopu and Adrian Munteanu",
year = "2019",
month = oct,
day = "3",
doi = "10.1049/el.2019.1757",
language = "English",
volume = "55",
pages = "1086–1088",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology",
number = "20",
}