Abstract
Semantic segmentation and deep learning methods have rarely been applied to fractional vegetation cover (FVC) segmentation tasks due to the lack of publicly available datasets for training deep learning models. FVC is a key indicator for assessing vegetation distribution, crop density, and crop responses to water availability and fertilizer application, yet conventional field-based measurement methods are time consuming, costly, labor intensive, and may lack the accuracy required for critical applications such as drought stress evaluation and water productivity. In this paper, we introduced causality-based deep learning techniques for FVC segmentation on a publicly available RGB dataset that consists of four ground cover crops: Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L. By separating causal from spurious correlations in pretrained features, using the stepwise intervention and reweighting (SIR) method at different encoder stages reduced confounding bias and enabled the models to learn more generalizable and task-relevant features. Extensive experiments on the FVC dataset, conducted with and without causality learning, showed that the proposed FCN + ResNet-50 model with causality learning and data augmentation achieved an accuracy of 94.80%, a precision of 94.97%, a recall of 94.35%, and an F1-score of 94.62%, which outperformed non-causal baselines and state-of-the-art transformer-based models including SegFormer and Mask2Former.
Abstract from the paper’s OpenAlex record; open copy licensed cc-by. Read at the source ↗
The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.
Claims 5
Claim · model-written summary
Stepwise intervention and reweighting reduced confounding bias and improved generalizable features.
- Taxon
- Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.
Cynodon dactylon
Oxalis stricta.JPG?width=60)
Phyla nodiflora
Frankenia thymifoliae1 - Outcome
- confounding biase0
- Direction
- reducede0
Not reportedComparatorValue
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
.JPG?width=500)
Verbatim
e0“using the stepwise intervention and reweighting (SIR) method at different encoder stages reduced confounding bias and enabled the models to learn more generalizable and task-relevant features”
e1“four ground cover crops: Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.”
Claim · model-written summary
The causal augmented model achieved 94.80% accuracy, outperforming comparison models.
- Taxon
- Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.Cynodon dactylonOxalis strictaPhyla nodifloraFrankenia thymifoliae1
- Outcome
- accuracye0
- Direction
- outperformede0
- Value
- 94.80%e0
- Comparator
- non-causal baselinese0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“the proposed FCN + ResNet-50 model with causality learning and data augmentation achieved an accuracy of 94.80%, a precision of 94.97%, a recall of 94.35%, and an F1-score of 94.62%, which outperformed non-causal baselines and state-of-the-art transformer-based models including SegFormer and Mask2Former”
e1“four ground cover crops: Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.”
Claim · model-written summary
The causal augmented model achieved 94.97% precision.
- Taxon
- Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.Cynodon dactylonOxalis strictaPhyla nodifloraFrankenia thymifoliae1
- Outcome
- precisione0
- Value
- 94.97%e0
Not reportedDirection
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“a precision of 94.97%”
e1“four ground cover crops: Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.”
Claim · model-written summary
The causal augmented model achieved 94.35% recall.
- Taxon
- Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.Cynodon dactylonOxalis strictaPhyla nodifloraFrankenia thymifoliae1
- Outcome
- recalle0
- Value
- 94.35%e0
Not reportedDirection
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“a recall of 94.35%”
e1“four ground cover crops: Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.”
Claim · model-written summary
The causal augmented model achieved 94.62% F1-score.
- Taxon
- Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.Cynodon dactylonOxalis strictaPhyla nodifloraFrankenia thymifoliae1
- Outcome
- F1-scoree0
- Value
- 94.62%e0
Not reportedDirection
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“an F1-score of 94.62%”
e1“four ground cover crops: Phyla nodiflora L., Cynodon dactylon, Frankenia thymifolia Desf., and Oxalis stricta L.”