Pith. sign in

REVIEW 4 major objections 6 minor 32 references

DV3+HED+: A DCNNs-based Framework to Monitor Temporary Works and ESAs in Railway Construction Project Using VHR Satellite Images

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read An edge-aware deep network labels temporary works and environmental sensitive areas in very-high-resolution satellite images of railway construction with over 80 percent overall accuracy.

desk verdict A useful applied integration of DV3+ and HED for railway construction monitoring, but the headline accuracy is not credible until the authors clarify whether the 15 test scenes were held out from the 572 training images. read the letter →

arxiv 1908.11080 v1 pith:GGTJK3CI submitted 2019-08-29 cs.CV

classification cs.CV
keywords railwayconstructiondeeplearningremotesensingconvolutionalneuralnetworksemanticsegmentationveryhighresolutionsatelliteimagesedgedetectionattentionloss
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that an integrated deep-learning pipeline can automatically map temporary construction works (borrow and spoil areas, camps, beam yards) and environmentally sensitive areas (resident houses) in very-high-resolution satellite images of railway construction. Its central contribution is a semantic segmentation network, DV3+HED+, that adds a holistically-nested edge detection sub-network and an attention cross-entropy loss to the DeepLabV3+ architecture. Tested on 15 QuickBird scenes along the Ruichang-Jiujiang railway from 2015 to 2017, the network reports overall accuracy above 80 percent in each year, consistently beating the DeepLabV3+ baseline and the version without the attention loss. If the result holds beyond this test site, routine satellite revisits could replace much of the manual visual inspection now needed for environmental and relocation compliance monitoring.

What carries the argument

The central object is the DV3+HED+ network itself. It chains two sub-networks: HED, a VGG-16-based holistically-nested edge detector that emits multi-scale side outputs fused into one boundary map, and DeepLabV3+, an encoder-decoder semantic segmentation network with atrous separable convolution. The two are joined by concatenating the boundary prediction with the original image as input to DeepLabV3+, and the whole model is trained jointly with an attention cross-entropy loss that upweights false negatives and false positives on boundary pixels. Boundary ground truth is generated automatically by applying a Sobel edge operator to the class-label polygons, avoiding manual boundary labelling.

What would settle it

Apply the same trained network to VHR satellite images of a different railway project with independently produced ground truth and compare overall and per-class accuracy; if the roughly 80 percent overall accuracy does not reproduce, or per-class accuracy for beam yards collapses, the paper's claim of a general monitoring framework is unsupported.

Watch

Extended reading notes

Core claim

The paper introduces DV3+HED+, a fully convolutional network that takes a true-color VHR image, first extracts a boundary map with the HED sub-network (trained on Sobel-derived boundary ground truth), then feeds the original image together with that boundary map into a DeepLabV3+ encoder-decoder for pixel-wise classification. The attention loss reweights the boundary detection so that hard, misclassified edge pixels contribute more to training, countering the class imbalance that makes ordinary cross-entropy blur object boundaries. On the Ruichang-Jiujiang railway data, reported overall accuracy is 80.05% (2015), 80.35% (2016), and 80.19% (2017), compared with 77.23/76.95/78.24% for DV3+ alone, and the paper describes these as clear improvements. The network also produces multi-temporal maps that reflect construction stages, including the disappearance of the beam yard after the project was completed in 2017.

Load-bearing premise

The reported accuracy depends on the 15 QuickBird scenes of the Ruichang-Jiujiang railway and their hand-checked labels being representative of railway construction sites generally; if the labels contain systematic errors or the test site is atypical, the results will not generalize.

Editorial extensions

If this is right

  • A single trained model can label borrow/spoil areas, camps, beam yards, and resident houses pixel-by-pixel across an entire railway corridor from very-high-resolution satellite images.
  • Repeated classification over time yields a change-detection timeline for construction activity, as shown by the reported peak in temporary works during 2016 and the removal of the beam yard after the 2017 completion.
  • The attention-loss recipe transfers to any semantic segmentation task with a strongly imbalanced boundary class, offering a way to sharpen object edges without extra manual annotation.
  • The framework can serve as a pre-screening step for environmental supervision, producing candidate maps that still need final human correction before official use.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the network architecture is not railway-specific, the same edge-plus-attention design could be retrained for other linear infrastructure such as highways, pipelines, or power lines; a cross-project accuracy study would test this directly.
  • A cross-sensor experiment (train on one satellite source, test on another) would isolate how much of the reported accuracy depends on the mix of QuickBird, GF-2, and Google Earth imagery in the training set.
  • The beam-yard class rests on only 25 ground-truth samples, so its per-class accuracy should be interpreted cautiously; a dedicated evaluation with more beam yards would reveal whether the edge-refinement gain holds for rare objects.
  • In practice, the segmentation maps could be paired with rule-based distance checks (for example, resident houses within 30 meters of the construction boundary) to generate automated relocation-compliance alerts.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The manuscript presents DV3+HED+, a semantic segmentation framework for monitoring temporary construction works (borrow/spoil areas, camps, beam yards) and environmentally sensitive areas (resident houses) along railways from very-high-resolution (VHR) satellite imagery. The network combines DeepLabV3+ with a HED boundary-detection subnetwork and an attention-based cross-entropy loss. The authors state that the model is trained on 572 true-color VHR images from QuickBird, GF-2, and Google Earth and tested on 15 QuickBird scenes of the Ruichang–Jiujiang railway. Table 5 reports overall accuracies of 80.05, 80.35 and 80.19% for 2015, 2016 and 2017, with the proposed method outperforming DV3+ and DV3+HED ablations. The paper also claims that the framework supports change detection and provides open-source code.

Significance. If the results are valid, this is a practically useful application of modern segmentation networks to an under-studied monitoring problem, with a sensible architectural combination (HED boundary supervision plus DeepLabV3+) and a clear attempt to address class imbalance through an attention loss. The explicit ablation structure and the plan to release code are strengths. However, the empirical evidence is currently too thin to support the advertised degree of certainty: only a single test corridor is used, the comparison is limited to ablations of the backbone, and there are no measures of variability or significance. The paper's contribution is therefore conditional on strengthening the evaluation and clarifying the data split.

major comments (4)
  1. [§3.1, Tables 2–3, and Abstract] Train/test disjointness is not established. Table 2, headed 'VHR images summary for training and testing', sums to exactly 572 images (86 QuickBird + 182 GF-2 + 304 Google Earth), which is precisely the number the abstract states as the training set. No row identifies the 15 Ruichang–Jiujiang QuickBird test scenes or explicitly states that they are excluded from those 572 images. Section 3.1 further says that GF-2 was used for 'ground truth labeling, training and testing' while QuickBird was only an auxiliary source for 'model training', which conflicts with the claim that testing was done on QuickBird imagery. If the 15 test scenes were part of the training data, every accuracy in Table 5 is invalid. The authors must state explicitly that the test scenes are disjoint from the training images and document the split (for example, by image IDs or date/location footprints).
  2. [§5.2, Table 5] The claimed improvements lack variability estimates and significance testing. In 2015, the gain of DV3+HED+ over DV3+HED is only +0.11 percentage points (80.05 vs. 79.94); in 2016 and 2017 the gains are +1.63 and +1.08. Without error bars, repeated runs, or statistical tests, the statement that the method has 'obvious improvements' is not substantiated. In addition, the comparison to the 'existing state-of-the-art approach' is only against the DV3+ backbone; no other established segmentation networks (e.g., U-Net, PSPNet, DeepLabV3, BiSeNet) are evaluated on this dataset, so the paper does not demonstrate superiority over current alternatives.
  3. [§5.2, Eq. (4)] The metric called 'accuracy' is defined as TP/(TP+FP), which is precision, not accuracy. Because the abstract and conclusions rely on the headline claim of 'overall accuracy of more than 80%', the authors must correct this definition and specify how the overall accuracy is computed (e.g., pixel-wise fraction of correctly classified pixels) separately from per-class precision and recall. The mislabeling makes it difficult to interpret the reported per-class numbers.
  4. [§4.1] The boundary ground truth for the HED subnetwork is generated by applying a Sobel edge detector to the class-label polygons. This is an unvalidated proxy: polygon boundaries may not coincide with true object boundaries in the imagery because of geometric correction errors, labeling inaccuracies, and the rasterization of vector shapes. Since the HED subnetwork and its attention loss are claimed as key components of the improvement, the paper should provide a quantitative evaluation of boundary detection quality; without it, the source of the reported improvement is unclear.
minor comments (6)
  1. [Tables 1 and 4] Table 4 labels borrow area, spoil area, and beam yard as 'ESA', whereas Table 1 classifies them as 'Temporary work'. The class taxonomy should be consistent throughout, and the borrow/spoil merge described in Section 5.2 should be reflected in the data description.
  2. [§3.2 and References] The text cites 'Guo[42]' for the data augmentation method, but the reference list contains only 37 entries and the relevant work appears to be reference [33]. The in-text citation needs to be corrected.
  3. [Abstract and Table 5] The abstract says 'more than 80%' while Table 5 reports 80.05, 80.35 and 80.19; it would be more precise to state the actual range or say 'about 80%'.
  4. [Throughout] There are several typographical and language issues: 'Ruichang-Jiujing' in the Conclusions, 'comparasion' in Figure 4's caption, 'interesting with our works' should be 'interested in our work', and 'serval years' should be 'several years'.
  5. [§4, Eq. (1)] Equation (1) is not typeset in the submitted text, and the variables in the loss relation are not fully defined. Please provide the complete equation and define all symbols.
  6. [§5.3] The hyperparameters of the attention loss are given only as α=4 and β=0.4; the authors should state how these values were chosen and whether the results are sensitive to them.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported accuracy is an empirical result on held-out QuickBird scenes, and the network components are drawn from external published work.

full rationale

No load-bearing circular step can be exhibited. The central empirical claim, that DV3+HED+ reaches an overall accuracy above 80 percent on 15 QuickBird test scenes, is evaluated on imagery that is described as a test set, and the numbers in Table 5 are not derived from the fitted network parameters or from any equation that re-inserts the training labels into the test metric. The architecture combines externally published components: DV3+ [7], HED [8], and the attention loss from DOOBNet [36]. The only self-citation, Guo [33]/[42], supplies an auxiliary selective-search data augmentation step; it does not define the test accuracy or the claimed improvement, so it is not load-bearing. The boundary ground truth is derived from the semantic labels with a Sobel operator, but this is an auxiliary supervision signal for the HED sub-network, not a definition of the final classification output; at test time the boundary map is predicted from the image. The possible overlap between the 15 test scenes and the 572 training images is a data-integrity concern rather than a circularity of the derivation, and the apparent citation-numbering inconsistency ([42] versus the reference list) is a referencing error, not a circular step. Therefore the paper is not circular in the sense assessed here.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim is an empirical accuracy result, so the ledger contains no invented entities. The hand-chosen attention loss parameters alpha=4 and beta=0.4 are free parameters; the unstated loss weighting coefficients are an additional tuning choice. The main assumptions are transfer learning from ImageNet, the adequacy of Sobel-derived boundary labels, the accuracy of manual or shapefile ground truth, and the merging of borrow and spoil areas into one class. No independent falsifiable entity is introduced.

free parameters (3)
  • alpha = 4
    Hand-chosen parameter in the attention loss (Eq. 3), reported in Section 5.3 as alpha=4; adjusts true positive vs false positive loss contributions and affects the boundary detection sub-network.
  • beta = 0.4
    Hand-chosen smoothing parameter in the attention loss (Section 5.3, beta=0.4). Chosen on the validation set, not derived from first principles.
  • loss weighting coefficients = not reported
    The total loss combines semantic segmentation cross entropy, HED side losses, fusion loss, and L2 regularization, but the paper does not specify the weighting coefficients; these are hand-chosen and affect the reported results.
assumptions (4)
  • domain assumption ImageNet-pretrained weights provide a useful initialization for VHR remote sensing semantic segmentation.
    Section 5.1 states that the authors fine-tune models that have been validated successfully in natural image semantic segmentation; the transferability of ImageNet features to satellite imagery is assumed from prior work (reference [27]) and not tested in this paper.
  • ad hoc to paper Sobel edge detection on class label polygons produces adequate boundary ground truth for the HED sub-network.
    Section 4.1: 'we use Sobel edge detection operator to generate the boundary ground truth data'; this avoids manual boundary labeling but assumes Sobel edges align with true object boundaries.
  • domain assumption The ESRI Shapefile reference data and visual interpretation provide accurate ground truth for training and testing.
    Section 3.1 and 3.2 describe ground truth derived from shapefiles and manual labeling; no label quality assessment is provided, so label errors would directly affect the reported accuracies.
  • domain assumption Borrow and spoil areas can be merged into one category because both present bare soil.
    Section 5.2: 'we take the borrow and spoil areas as a same target object category'; this is a domain decision that reduces the number of classes and shapes the per-class accuracy numbers.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DV3+HED+: A DCNNs-based Framework to Monitor Temporary Works and ESAs in Railway Construction Project Using VHR Satellite Images." pith.science (2026). https://pith.science/paper/GGTJK3CI

@misc{pith2026190811080,
  author       = {Pith},
  title        = {Pith review of: DV3+HED+: A DCNNs-based Framework to Monitor Temporary Works and ESAs in Railway Construction Project Using VHR Satellite Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GGTJK3CI}},
  note         = {Machine review of arXiv:1908.11080}
}
read the original abstract

Current VHR(Very High Resolution) satellite images enable the detailed monitoring of the earth and can capture the ongoing works of railway construction. In this paper, we present an integrated framework applied to monitoring the railway construction in China, using QuickBird, GF-2 and Google Earth VHR satellite images. We also construct a novel DCNNs-based (Deep Convolutional Neural Networks) semantic segmentation network to label the temporary works such as borrow & spoil area, camp, beam yard and ESAs(Environmental Sensitive Areas) such as resident houses throughout the whole railway construction project using VHR satellite images. In addition, we employ HED edge detection sub-network to refine the boundary details and attention cross entropy loss function to fit the sample class disequilibrium problem. Our semantic segmentation network is trained on 572 VHR true color images, and tested on the 15 QuickBird true color images along Ruichang-Jiujiang railway during 2015-2017. The experiment results show that compared with the existing state-of-the-art approach, our approach has obvious improvements with an overall accuracy of more than 80%.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 24 canonical work pages

  1. [1]

    Support vector machines in remote sensing: A review

    Mountrakis, G.; Im, J.; Ogole, C. Support vector machines in remote sensing: A review. ISPRS J. Photogramm. Remote Sens . 2011 , 66 , 247 –

  2. [2]

    Neural network classification of remote - sensing data

    Miller, D.M.; Kaminsky, E.J.; Rana, S. Neural network classification of remote - sensing data. Comput. Geosci . 1995 , 21 , 377 –

  3. [3]

    Random forest classifier for remote sensing classification

    Pal, M. Random forest classifier for remote sensing classification. Int. J. Remote Sens . 2005 , 26 , 217 –

  4. [4]

    Object - b ased image analysis for remote sensing

    Blaschke , T. Object - b ased image analysis for remote sensing . Object - based image analysis for remote sensing . 2010 , 65 (1) , 2 - 16

  5. [5]

    Object Recognition from Local gre Scale - Invariant Features

    Lowe, D.G. Object Recognition from Local gre Scale - Invariant Features. Proc. Int. Conf. Comput. Vis. 1999 , 2 , 1150 –

  6. [6]

    Histog rams of Oriented Gradients for Human Detection

    Dalal, N.; Triggs, B. Histog rams of Oriented Gradients for Human Detection. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. 2005 , 1 , 886 –

  7. [7]

    C., Zhu, Y., Papandreou, G., Schroff, F., Adam, H

    Chen, L. C., Zhu, Y., Papandreou, G., Schroff, F., Adam, H. Encoder - decoder with atrous separable convolution for semantic image segmen tation. 2018 , arXiv:1802.02611v3

  8. [8]

    Holistically - Nested Edge Detection[J]

    Xie S, Tu Z. Holistically - Nested Edge Detection[J]. International Journal of Computer Vision, 2015, 125(1 - 3):3 -

Show all 32 references
  1. [9]

    Wickramasinghe , D. C. , Vu , T.T., Maul , T. Satellite remote - sensing monitoring of a railway construction project. International Journal of Remote Sensing , 2017 , 39 ( 6 ) , 1754 – 1769

  2. [10]

    , Ferretti , A

    Giannico , C. , Ferretti , A. , Alberti , S . Application of satellite radar interferometry for tunnel and underground infrastructures damage assessment and monitoring . In Lif e - Cycle and Sustainability of Civil Infrastructure Systems: Proceedings of the Third International...

  3. [11]

    M onitoring abandoned dreg fields of high - speed railway construction with UAV remote sensing technology

    Lin, J., Wang, Z., Wang, Y., Lin, Y., Du, X . M onitoring abandoned dreg fields of high - speed railway construction with UAV remote sensing technology. In: International Conference on Intelligent Earth Observing and Applications 2015 . International Society for Optics and Photonics

  4. [12]

    Chang, L., Dollevoet, R. P. B. J., Hanssen, R. F. Nationwide railway monitoring using satellite SAR interferometry. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 1 -

  5. [13]

    Automated recognition of railroad infrastructure in r ural areas from lidar data

    Arastounia, M. Automated recognition of railroad infrastructure in r ural areas from lidar data. Remote Sensing, 2015 , 7(11), 14916 - 14938

  6. [14]

    Fully convolutional networks for semantic segmentation

    Long, J.; Shelhamer, E.; Darrell, T. Fully convolutional networks for semantic segmentation. In Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7 – 12 June

  7. [15]

    The pascal visual object classes challenge: A retrospective

    Everingham, M.; Eslami, S.A.; Van Go ol, L.; Williams, C.K.; Winn, J.; Zisserman, A. The pascal visual object classes challenge: A retrospective. Int. J. Comput. Vis . 2015 , 111 , 98 –

  8. [16]

    - Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Dollar, P.; Zitnick, C.L

    Lin, T. - Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Dollar, P.; Zitnick, C.L. Microsoft coco: Common objects in context. In Proceedings of the European Conference on Computer Vision ECCV, Zurich, Switzerland, 6 – 12 September 2014; pp. 740 –

  9. [17]

    SegNet: A Deep Convolutional Encoder - Decoder Architecture for Im age Segmentation

    Vijay, B.; Kendall, A.; Cipolla, R. SegNet: A Deep Convolutional Encoder - Decoder Architecture for Im age Segmentation. arXiv 2015 , arXiv:1511.00561

  10. [18]

    U - Net: Convolutional Networks for Biomedical Image Segmentation

    Ronneberger, O.; Fischer, P.; Brox, T. U - Net: Convolutional Networks for Biomedical Image Segmentation. In Proceedings of the International Conference on Medical image computing and computer - assisted interven tion, Munich, Germany, 5 – 9 October 2015; pp. 234 – 241

  11. [19]

    RefineNet: Multi - Path Refinement Networks for High - Resolution Semantic Segmentation

    Lin, G.; Milan, A.; Shen, C.; Reid, I. RefineNet: Multi - Path Refinement Networks for High - Resolution Semantic Segmentation. arXiv 2016 , arXiv:1611.06612

  12. [21]

    Linknet: exploiting encoder representations for efficient semantic segmentation

    Chaurasia, A., & Culurciello, E. Linknet: exploiting encoder representations for efficient semantic segmentation. arXiv 2017 , arXiv: 1707.03718

  13. [22]

    Pyramid Scene Parsing Network

    Zhao, H.; Shi, J.; Qi, X.; Wang, X.; Jia, J. Pyramid Scene Parsing Network. arXiv 2016 , arXiv:1612.01105

  14. [24]

    - C.; Papandreou, G.; K okkinos, I.; Murphy, K.; Yuille, A.L

    Chen, L. - C.; Papandreou, G.; K okkinos, I.; Murphy, K.; Yuille, A.L. Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs. arXiv 2014 , arXiv:1412.7062v1

  15. [25]

    - C.; Papandreou, G.; Kokkinos, I.; Murphy, K.; Yuille, A.L

    Chen, L. - C.; Papandreou, G.; Kokkinos, I.; Murphy, K.; Yuille, A.L. DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs. arXiv 2016 , arXiv:1606.00915v1

  16. [26]

    C., Papandreou, G., Schroff, F., A dam, H

    Chen, L. C., Papandreou, G., Schroff, F., A dam, H. Rethinking atrous convolution for semantic image segmentation. arXiv 201 7 , arXiv: 1706.05587

  17. [28]

    Dense semantic labeling of sub - decimeter resolution images with convolutional neu ral networks

    Volpi, M.; Tuia, D. Dense semantic labeling of sub - decimeter resolution images with convolutional neu ral networks. IEEE Trans. Geosci. Remote Sen. 2017 , 55 , 881 –

  18. [29]

    Hourglass - ShapeNetwork Based Semantic Segmentation for High Resolution Aerial Imagery

    Liu, Y.; Nguyen, D.; Deligiannis, N.; Ding, W.; Munteanu, A. Hourglass - ShapeNetwork Based Semantic Segmentation for High Resolution Aerial Imagery. Remote Sens. 2017 , 9 , 522

  19. [30]

    High - resolution semantic labeling with convolutional neural networks

    Maggiori, E.; Tarabalka, Y.; Charpiat, G.; Alliez, P. High - resolution semantic labeling with convolutional neural networks. arXiv 2016 , arXiv:1611.01962

  20. [31]

    Building Footprint Extraction from Hig h - Resolution Images via Spatial Residual Inception Convolutional Neural Network

    Liu, P.; Liu, X.; Liu, M.; Shi, Q.; Yang, J.; Xu, X.; Zhang, Y. Building Footprint Extraction from Hig h - Resolution Images via Spatial Residual Inception Convolutional Neural Network. Remote Sens. 2019, 11,

  21. [32]

    Classification with an edge: Improving semantic image segmentation w ith boundary detection

    Marmanis, D.; Schindler, K.; Wegner, J.D.; Galliani, S.; Datcu, M.; Stilla, U. Classification with an edge: Improving semantic image segmentation w ith boundary detection. ISPRS Journal of Photogrammetry and Remote Sensing, 135, 2018, 158 -

  22. [33]

    Pixel - Wise Classification Method for High Resolution Remote Sensing Imagery Usi ng Deep Neural Networks

    Guo, R.; Liu, J.; Li, N.; Liu, S.; Chen, F.; Cheng, B.; Duan, J.; Li, X.; Ma, C. Pixel - Wise Classification Method for High Resolution Remote Sensing Imagery Usi ng Deep Neural Networks. ISPRS Int. J. Geo - Inf. 2018, 7,

  23. [35]

    A Y - Net deep learning method for road seg mentation using high - resolution visible remote sensing images

    Li, Y., Xu, L., Rao, J., Guo, L., Yan, Z., & Jin, S. A Y - Net deep learning method for road seg mentation using high - resolution visible remote sensing images. Remote sensing letters, 2019, 10(4), 381 -

  24. [37]

    Very Deep Convoluti onal Networks for Large - Scale Image Recognition

    Simonyan, K.; Zisserman, A. Very Deep Convoluti onal Networks for Large - Scale Image Recognition. In Proceedings of the 3rd International Conference on Learning Representations, San Diego, CA, USA, 7 – 9 May 2015

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.