Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Interpreting a Semantic Segmentation Model for Coastline Detection

Conor O'Sullivan, Seamus Coveney, Xavier Monteys, Soumyabrata Dev

TL;DR

A deep-learning semantic segmentation model used to classify coastline satellite images into land and water is interpreted to build trust in the model and gain new insight into the process of coastal water body extraction, and which spectral bands are important for predicting segmentation masks are sought.

Abstract

We interpret a deep-learning semantic segmentation model used to classify coastline satellite images into land and water. This is to build trust in the model and gain new insight into the process of coastal water body extraction. Specifically, we seek to understand which spectral bands are important for predicting segmentation masks. This is done using a permutation importance approach. Results show that the NIR is the most important spectral band. Permuting this band lead to a decrease in accuracy of 38.12 percentage points. This is followed by Water Vapour, SWIR 1, and Blue bands with 2.58, 0.78 and 0.19 respectively. Water Vapour is not typically used in water indices and these results suggest it may be useful for water body extraction. Permuting, the Coastal Aerosol, Green, Red, RE1, RE2, RE3, RE4, and SWIR 2 bands did not decrease accuracy. This suggests they could be excluded from future model builds reducing complexity and computational requirements.

Interpreting a Semantic Segmentation Model for Coastline Detection

TL;DR

A deep-learning semantic segmentation model used to classify coastline satellite images into land and water is interpreted to build trust in the model and gain new insight into the process of coastal water body extraction, and which spectral bands are important for predicting segmentation masks are sought.

Abstract

We interpret a deep-learning semantic segmentation model used to classify coastline satellite images into land and water. This is to build trust in the model and gain new insight into the process of coastal water body extraction. Specifically, we seek to understand which spectral bands are important for predicting segmentation masks. This is done using a permutation importance approach. Results show that the NIR is the most important spectral band. Permuting this band lead to a decrease in accuracy of 38.12 percentage points. This is followed by Water Vapour, SWIR 1, and Blue bands with 2.58, 0.78 and 0.19 respectively. Water Vapour is not typically used in water indices and these results suggest it may be useful for water body extraction. Permuting, the Coastal Aerosol, Green, Red, RE1, RE2, RE3, RE4, and SWIR 2 bands did not decrease accuracy. This suggests they could be excluded from future model builds reducing complexity and computational requirements.
Paper Structure (2 sections, 5 equations, 3 figures, 3 tables)

This paper contains 2 sections, 5 equations, 3 figures, 3 tables.

Table of Contents

  1. Appendix
  2. Remove images

Figures (3)

  • Figure 1: Model architecture consisting of an encoder, bottleneck and decoder. Skip connections are used to concatenate layers in the encoder to symmetrical layers in the decoder. The number below each layer is the number of channels in that layer.
  • Figure 2: Permutation importance scores for spectral bands used as input into the coastal segmentation model. The scores give a decrease in average accuracy when the respective band is permuted.
  • Figure 3: Permutation importance scores for coastline segmentation model. The scores give the percentage change in average accuracy when combinations are bands are permuted.