Advanced search
1 file | 9.33 MB Add to list

Efficient training procedures for multi-spectral demosaicing

Ivana Shopovska (UGent) , Ljubomir Jovanov (UGent) and Wilfried Philips (UGent)
(2020) SENSORS. 20(10). p.2850-2873
Author
Organization
Abstract
The simultaneous acquisition of multi-spectral images on a single sensor can be efficiently performed by single shot capture using a mutli-spectral filter array. This paper focused on the demosaicing of color and near-infrared bands and relied on a convolutional neural network (CNN). To train the deep learning model robustly and accurately, it is necessary to provide enough training data, with sufficient variability. We focused on the design of an efficient training procedure by discovering an optimal training dataset. We propose two data selection strategies, motivated by slightly different concepts. The general term that will be used for the proposed models trained using data selection is data selection-based multi-spectral demosaicing (DSMD). The first idea is clustering-based data selection (DSMD-C), with the goal to discover a representative subset with a high variance so as to train a robust model. The second is an adaptive-based data selection (DSMD-A), a self-guided approach that selects new data based on the current model accuracy. We performed a controlled experimental evaluation of the proposed training strategies and the results show that a careful selection of data does benefit the speed and accuracy of training. We are still able to achieve high reconstruction accuracy with a lightweight model.
Keywords
Electrical and Electronic Engineering, Image Processing, Deep Learning, Computer Vision, RGB, NIR, multispectral, demosaicing, deep learning, data sampling, active learning, IMAGE, ALGORITHM

Downloads

  • sensors-20-02850-v3 FinalPublished.pdf
    • full text (Published version)
    • |
    • open access
    • |
    • PDF
    • |
    • 9.33 MB

Citation

Please use this url to cite or link to this publication:

MLA
Shopovska, Ivana, et al. “Efficient Training Procedures for Multi-Spectral Demosaicing.” SENSORS, vol. 20, no. 10, 2020, pp. 2850–73, doi:10.3390/s20102850.
APA
Shopovska, I., Jovanov, L., & Philips, W. (2020). Efficient training procedures for multi-spectral demosaicing. SENSORS, 20(10), 2850–2873. https://doi.org/10.3390/s20102850
Chicago author-date
Shopovska, Ivana, Ljubomir Jovanov, and Wilfried Philips. 2020. “Efficient Training Procedures for Multi-Spectral Demosaicing.” SENSORS 20 (10): 2850–73. https://doi.org/10.3390/s20102850.
Chicago author-date (all authors)
Shopovska, Ivana, Ljubomir Jovanov, and Wilfried Philips. 2020. “Efficient Training Procedures for Multi-Spectral Demosaicing.” SENSORS 20 (10): 2850–2873. doi:10.3390/s20102850.
Vancouver
1.
Shopovska I, Jovanov L, Philips W. Efficient training procedures for multi-spectral demosaicing. SENSORS. 2020;20(10):2850–73.
IEEE
[1]
I. Shopovska, L. Jovanov, and W. Philips, “Efficient training procedures for multi-spectral demosaicing,” SENSORS, vol. 20, no. 10, pp. 2850–2873, 2020.
@article{8663961,
  abstract     = {The simultaneous acquisition of multi-spectral images on a single sensor can be efficiently performed by single shot capture using a mutli-spectral filter array. This paper focused on the demosaicing of color and near-infrared bands and relied on a convolutional neural network (CNN). To train the deep learning model robustly and accurately, it is necessary to provide enough training data, with sufficient variability. We focused on the design of an efficient training procedure by discovering an optimal training dataset. We propose two data selection strategies, motivated by slightly different concepts. The general term that will be used for the proposed models trained using data selection is data selection-based multi-spectral demosaicing (DSMD). The first idea is clustering-based data selection (DSMD-C), with the goal to discover a representative subset with a high variance so as to train a robust model. The second is an adaptive-based data selection (DSMD-A), a self-guided approach that selects new data based on the current model accuracy. We performed a controlled experimental evaluation of the proposed training strategies and the results show that a careful selection of data does benefit the speed and accuracy of training. We are still able to achieve high reconstruction accuracy with a lightweight model.},
  author       = {Shopovska, Ivana and Jovanov, Ljubomir and Philips, Wilfried},
  issn         = {1424-8220},
  journal      = {SENSORS},
  keywords     = {Electrical and Electronic Engineering,Image Processing,Deep Learning,Computer Vision,RGB,NIR,multispectral,demosaicing,deep learning,data sampling,active learning,IMAGE,ALGORITHM},
  language     = {eng},
  number       = {10},
  pages        = {2850--2873},
  title        = {Efficient training procedures for multi-spectral demosaicing},
  url          = {http://dx.doi.org/10.3390/s20102850},
  volume       = {20},
  year         = {2020},
}

Altmetric
View in Altmetric
Web of Science
Times cited: