Pith. sign in

REVIEW 4 major objections 4 minor 49 references

MRIFE: A Mask-Recovering and Interactive-Feature-Enhancing Semantic Segmentation Network For Relic Landslide Detection

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that a dual-branch network that reconstructs masked landslide-edge features and contrasts them against background features, trained with self-distillation, substantially improves relic landslide segmentation in small…

desk verdict Plausible architecture, but the reported landslide-IoU gain is not trustworthy as reported — internal numeric inconsistency and likely selection-on-test. read the letter →

arxiv 2411.17167 v1 pith:N4RT7F4O submitted 2024-11-26 cs.CV

Juefei He , Yuexing Peng , Wei Li , Junchuan Yu , Daqing Ge , Wei Xiang This is my paper · ORCID
classification cs.CV
keywords reliclandslidedetectionsemanticsegmentationhigh-resolutionremotesensingimagerymaskedfeaturemodelingsupervisedcontrastivelearningself-distillationdual-branchnetworkvisualblur
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

Relic landslides—old, stable slides that can reactivate—are hard to spot in high-resolution satellite images because their edges have blurred into the surrounding landscape, and good training data are scarce. The paper argues that both problems can be attacked at once by adding a second, self-distilled feature-enhancement branch to a standard semantic segmentation network. That branch masks out landslide-edge and background blocks, reconstructs the masked features, and uses supervised contrastive learning to push edge features away from background features in the learned representation. On a real relic landslide dataset, the resulting MRIFE model raises landslide IoU from 0.338 to 0.393 and precision from 0.423 to 0.535 relative to the DeepLabv3+ baseline, while also reducing early overfitting.

What carries the argument

The load-bearing mechanism is a dual-branch interactive architecture. The mask module splits each 512x512 image into 8x8 pixel blocks (each corresponding to one feature point at stride 8) and labels them as non-landslide (fewer than 7 landslide pixels), landslide edge (7 to 57), or landslide interior (more than 57). It discards interior blocks and randomly masks equal numbers of edge and background blocks. Masked feature modeling reconstructs the masked feature points from the teacher network's features of the unmasked image using an MSE loss, while semantic feature contrast enhancement applies a supervised contrastive loss to pull together edge-block features and push apart edge versus background features across two different samples. A teacher-student network with exponential-moving-average updates provides self-distillation, and the enhanced features are added point-to-point to the segmentation branch's features before a shared decoder produces the final mask.

What would settle it

Run MRIFE on an independent relic landslide dataset from a different region, with its own expert labels; if landslide IoU does not improve over the same baselines by a comparable margin, the central claim fails. A cheaper check: have a second set of experts re-delineate boundaries and see whether the reported precision gain survives label noise.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that visual blur in relic landslide segmentation can be treated as a feature-separation problem rather than purely a representation-capacity problem. MRIFE couples a segmentation branch with a feature enhancement branch that deliberately masks the most informative image regions—landslide edge blocks and background blocks, defined by landslide-pixel counts per 8x8 block—then learns to reconstruct those masked feature points at the feature-map level and to contrast edge features against background features across two different samples. A teacher-student self-distillation scheme stabilises this multi-task training and mitigates overfitting on the small dataset. The result is a segmentation model whose predictions concentrate on landslide boundaries (back walls and side walls) and whose quantitative performance on the test set improves over both the baseline and the authors' previous model, with precision increasing from 0.4226 to 0.5347, landslide IoU from 0.3381 to 0.3934, mIoU from 0.6405 to 0.6680, and F1 from 0.5054 to 0.5646.

Load-bearing premise

The measured gains assume the expert-drawn landslide boundaries in the single study-region dataset are accurate and representative; if those labels carry systematic boundary errors or the region does not generalize, the improvements will not transfer.

Editorial extensions

If this is right

  • Relic landslide segmentation on small high-resolution remote sensing datasets can be improved without external data by supervising feature reconstruction and contrast at the mask-block level.
  • The feature enhancement branch, not just the segmentation backbone, drives the precision gain: adding masked feature modeling raises landslide IoU from 0.338 to 0.366, and adding the contrastive task raises it further to 0.393.
  • Self-distillation demonstrably curbs early overfitting in the small-data regime, as the training-loss curves for the model without it diverge early.
  • Masking only edge and background blocks beats masking interior blocks or all blocks, indicating that landslide interiors carry little discriminative signal in visually blurred relics.
  • Point-to-point addition of the two branches' features yields better fusion than channel concatenation with learned weighting.

Reading between the lines

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

  • The same recipe—masking label-identified transition zones and contrasting them against background—could transfer to other remote-sensing targets that are visually subtle, such as ancient agricultural terraces, archaeological mounds, or slow-moving faults, provided per-pixel labels exist.
  • The reported gains are precision-heavy: recall falls from 0.6284 (baseline) to 0.5981 (MRIFE), so the improvement is a rebalancing toward fewer false positives, and applications that prioritise recall would need threshold or loss re-weighting.
  • Because all experiments use one study region and one expert labelling procedure, the method's generalisation to different geomorphic settings is an open question; a multi-region evaluation with independent labels would be the natural next test.
  • The block-masking design is tied to the encoder's stride-8 downsampling; porting MRIFE to a transformer backbone or a different stride would require re-deriving the block-to-feature-point mapping.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes MRIFE, a dual-branch semantic segmentation network for relic landslide detection in high-resolution remote sensing images. The method combines a segmentation branch with a feature-enhancement branch that performs masked feature modeling (MFM), supervised contrastive learning (SFCE), and self-distillation via a teacher-student EMA architecture. The authors report that MRIFE improves precision, mIoU, landslide IoU, and F1-score over Deeplabv3+ and their prior ICSSN on a private relic landslide dataset, and they include ablation studies, Grad-CAM visualizations, and 'cross-validation' experiments to justify design choices such as mask block size, mask feature strategy, and fusion method.

Significance. If the reported improvements are reliable, MRIFE addresses a genuine and difficult problem: visually ambiguous relic landslides combined with small training datasets. The central idea of selectively masking edge and background blocks and performing cross-sample contrastive learning is interesting, and the authors provide a useful ablation showing the contribution of each component. The paper also presents qualitative evidence via Grad-CAM that the feature-enhancement branch focuses on landslide boundaries, which aligns with the intended mechanism. However, the empirical support is currently undermined by internal inconsistencies in the reported numbers, an apparent selection-on-test procedure, and the absence of uncertainty quantification. The method is plausible and the weaknesses are addressable, but the evidence as presented does not yet establish the claimed 5.5-point landslide-IoU improvement.

major comments (4)
  1. [Abstract and Section V-A vs. Table III] The reported landslide IoU for the final MRIFE model is inconsistent: the abstract and Section V-A state 0.3934, while Table III lists 0.3975 for the same configuration. Tables V and VI also report 0.3934 for the selected mask size and feature strategy. This discrepancy must be reconciled before the results can be considered reproducible.
  2. [Section V-C, Tables V and VI] The three 'cross-validation' experiments select the mask pixel size, mask feature strategy, and fusion method, but the paper never states that these selections were made on the validation split rather than the test split. Given the dataset division in Table I (60 test slide images), if the same test set was used for both model selection and final evaluation, the reported improvements are optimistically biased. The authors should either use a separate validation split for all hyperparameter choices or perform nested cross-validation, and report only the final test metrics.
  3. [Section IV-A, Table I, Section V-A] The evaluation rests on a single train/validation/test split of a private dataset with only 60 positive test images, and no error bars, confidence intervals, or repeated runs are reported. The observed absolute landslide-IoU gain of 0.055 could easily fall within random fluctuation for a test set of this size. The authors should report results over multiple random seeds or a k-fold cross-validation, with mean and standard deviation, to support the claim that the improvement is not due to chance.
  4. [Section III-C4, Eq. (5)] Equation (5) defines the indicator function I[i≠j] as 'value of 1 for i = j otherwise 0 for i ≠ j'. With this definition, the sum over j of I[i≠j]·log(·) is identically zero, making L_sup_i zero for every i. This contradicts the intended supervised contrastive loss and means the SFCE loss as written is non-functional. The definition of the indicator or the summation index must be corrected.
minor comments (4)
  1. [Section IV-C, Eq. (12)] The formula labeled '1 − mIoU' is actually the landslide-class IoU (TP/(TP+FP+FN)), not one minus the mean IoU. The notation is confusing and should be renamed, e.g., 'landslide IoU' or '1-IoU' as used elsewhere.
  2. [Section I and Section V-B] The contribution text describes a '5.5% performance improvement' in landslide IoU, but the numbers (0.3381 to 0.3934) represent an increase of 5.5 percentage points, not 5.5%. The same wording issue appears in Section V-B where '11%' and '6%' are percentage-point changes.
  3. [Section V-C] The subsection title 'Cross-validation' is misleading: the three experiments are model-selection procedures on a single split, not cross-validation in the statistical sense. A more accurate title would be 'Design-Choice Analysis' or 'Hyperparameter Selection'.
  4. [Section V-B, Fig. 6] The self-distillation ablation is supported only by training loss curves, without quantitative validation metrics. Reporting final validation or test metrics for the with- and without-self-distillation variants would make the claim more concrete.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: MRIFE's gains are empirical comparisons, not derived from its own definitions; self-citation to prior ICSSN is not load-bearing.

full rationale

This is an empirical benchmarking paper rather than a derivation, so most circularity patterns do not apply. The only self-reference is the authors' prior ICSSN (ref. [40]), which supplies the encoder/decoder components and serves as a comparison baseline; because that model is an external published system and the comparison is run on the same private dataset, its reuse does not make the reported MRIFE gains definitional. Section III-C1's statement that landslide-interior blocks 'contribute minimally' is not assumed; Section V-C explicitly tests the selection of mask features and reports that the edge-and-background strategy outperforms center-and-background and all-feature strategies. The Section V-C hyperparameter/configuration choices (mask size, mask feature strategy, fusion method) are selected using reported performance metrics; the paper does not state whether these selection runs used the separate validation split or the test split, and the abstract's headline 1-IoU of 0.3934 differs from Table III's 0.3975 for the same model. Those are evaluation-protocol and reporting-consistency concerns, not circularity: no equation, fitted parameter, or self-citation forces the claimed improvement to equal its own input. Under the hard rules requiring an explicit reduction, no circular step is established.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a private dataset and on several hyperparameters and design choices whose values are either unspecified (alpha, beta, gamma, tau) or chosen by performance on the same benchmark (mask size, feature selection, fusion method). No new physical entities are introduced.

free parameters (6)
  • alpha, beta, gamma = not specified
    Loss weights in the joint loss Eq. (8); values are tuned by hand but never reported.
  • tau = not specified
    Temperature in the supervised contrastive loss Eq. (5); value not given.
  • lambda_ema = 0.996
    EMA momentum for the teacher network in Eq. (1).
  • mask_thresholds = 7, 57 pixels
    Thresholds classifying 8x8 blocks into background, edge, and interior in Section III-C1; chosen by design, not derived.
  • mask_block_size = 8x8
    Selected from 8x8 vs 16x16 in Section V-C1 based on test performance.
  • mask_feature_strategy = edge and background
    Selected from three strategies in Section V-C2 based on test performance.
assumptions (4)
  • domain assumption Expert-labeled ground truth in the private dataset is accurate and reliable.
    The dataset is labeled by experts following the procedure in Section IV-A, but no inter-observer agreement or independent verification is reported.
  • domain assumption The private dataset is representative of relic landslide detection from HRSI.
    The study area is one region in northwest China; no external dataset is used to validate generalization.
  • ad hoc to paper Interior landslide blocks contribute minimally and can be discarded.
    Stated in Section III-C1 and 'validated' by a single comparison in Section V-C2 on the same dataset.
  • domain assumption The encoder/decoder from prior work [40] are suitable and their effectiveness is taken as given.
    The segmentation branch is based on the authors' previous ICSSN; no independent benchmark is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MRIFE: A Mask-Recovering and Interactive-Feature-Enhancing Semantic Segmentation Network For Relic Landslide Detection." pith.science (2026). https://pith.science/paper/N4RT7F4O

@misc{pith2026241117167,
  author       = {Pith},
  title        = {Pith review of: MRIFE: A Mask-Recovering and Interactive-Feature-Enhancing Semantic Segmentation Network For Relic Landslide Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N4RT7F4O}},
  note         = {Machine review of arXiv:2411.17167}
}
read the original abstract

Relic landslide, formed over a long period, possess the potential for reactivation, making them a hazardous geological phenomenon. While reliable relic landslide detection benefits the effective monitoring and prevention of landslide disaster, semantic segmentation using high-resolution remote sensing images for relic landslides faces many challenges, including the object visual blur problem, due to the changes of appearance caused by prolonged natural evolution and human activities, and the small-sized dataset problem, due to difficulty in recognizing and labelling the samples. To address these challenges, a semantic segmentation model, termed mask-recovering and interactive-feature-enhancing (MRIFE), is proposed for more efficient feature extraction and separation. Specifically, a contrastive learning and mask reconstruction method with locally significant feature enhancement is proposed to improve the ability to distinguish between the target and background and represent landslide semantic features. Meanwhile, a dual-branch interactive feature enhancement architecture is used to enrich the extracted features and address the issue of visual ambiguity. Self-distillation learning is introduced to leverage the feature diversity both within and between samples for contrastive learning, improving sample utilization, accelerating model convergence, and effectively addressing the problem of the small-sized dataset. The proposed MRIFE is evaluated on a real relic landslide dataset, and experimental results show that it greatly improves the performance of relic landslide detection. For the semantic segmentation task, compared to the baseline, the precision increases from 0.4226 to 0.5347, the mean intersection over union (IoU) increases from 0.6405 to 0.6680, the landslide IoU increases from 0.3381 to 0.3934, and the F1-score increases from 0.5054 to 0.5646.

Figures

Figures reproduced from arXiv: 2411.17167 by the authors.

Figure 1
Figure 1. Framework of the MRIFE. The two branches independently extract features and the decoder recovers the input images after feature fusion. A specially [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. (a) Original sample with 512×512 pixels; (b) part of original sample with 64 × 64 pixels; (c)-(e) masked block with 8 × 8 pixels: non-landslide block, landslide edge block, and landslide interior block. 1) Mask Module: We partition the input image of size 512 × 512 into 64 × 64 blocks, each block consisting of 8 × 8 pixels. Since the downsampling rate of the feature extractor is 8, each 8 × 8 block in the input imag… view at source ↗
Figure 3
Figure 3. (a) Landslide similar to the background; (b) label; (c)-(e) side view/top [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Visualized results of the comparative experiments. (a) Input; (b) label; (c)-(e) prediction: baseline, ICSSN, and MRIFE. The red line indicates the [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Grad-CAM results of the comparative experiments. (a) Input; (b) label; (c)-(e) heat map: baseline, ICSSN, and MRIFE. Red line indicates the boundary. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 7
Figure 7. Figure 7: Grad-CAM results. (a) Input; (b) label; (c)-(e) heat map: MRIFE, [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 6
Figure 6. Figure 6: Loss curve. (a) Without self-distillation framework; (b) with self [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 8
Figure 8. Figure 8: Grad-CAM results. (a) Input; (b) label; (c)-(e) [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 35 canonical work pages

  1. [1]

    Reactivation mechanism of old landslide triggered by coupling of fault creep and water infiltration: a case study from the east tibetan plateau,

    Y . Zhang, S. Ren, X. Liu, C. Guo, J. Li, J. Bi, and L. Ran, “Reactivation mechanism of old landslide triggered by coupling of fault creep and water infiltration: a case study from the east tibetan plateau,” Bulletin of Engineering Geology and the Environment , vol. 82, no. 8, p. 291, 2023

    work page 2023

  2. [2]

    Detection of landslide using high resolution satellite data and analysis using entropy,

    I. Shaik, S. Kameswara Rao, and B. Penta, “Detection of landslide using high resolution satellite data and analysis using entropy,” in Proceedings of International Conference on Remote Sensing for Disaster Manage- ment: Issues and Challenges in Disaster Management . Springer, 2019, pp. 243–250

    work page 2019

  3. [3]

    The 22 march 2014 oso landslide, washington, usa,

    J. Wartman, D. R. Montgomery, S. A. Anderson, J. R. Keaton, J. Beno ˆıt, J. dela Chapelle, and R. Gilbert, “The 22 march 2014 oso landslide, washington, usa,” Geomorphology, vol. 253, pp. 275–288, 2016

    work page 2014

  4. [4]

    Hu, “The historical transformation of the landslid- ing causes and factors in the border slopes of loessial JOURNAL OF LATEX CLASS FILES, VOL

    G. Hu, “The historical transformation of the landslid- ing causes and factors in the border slopes of loessial JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 11 highland in the baoji-changxing area,” Journal of Xi’an College of Geology , vol. 8, no. 4, pp. 23–27, 1986

    work page 2021

  5. [5]

    Landslide inventory maps: New tools for an old problem,

    F. Guzzetti, A. C. Mondini, M. Cardinali, F. Fiorucci, M. Santangelo, and K.-T. Chang, “Landslide inventory maps: New tools for an old problem,” Earth-Science Reviews, vol. 112, no. 1-2, pp. 42–66, 2012

    work page 2012

  6. [6]

    Preparation of earthquake-triggered landslide in- ventory maps using remote sensing and gis technologies: Principles and case studies,

    C. Xu, “Preparation of earthquake-triggered landslide in- ventory maps using remote sensing and gis technologies: Principles and case studies,” Geoscience Frontiers, vol. 6, no. 6, pp. 825–836, 2015

    work page 2015

  7. [7]

    Comparing landslide inventory maps,

    M. Galli, F. Ardizzone, M. Cardinali, F. Guzzetti, and P. Reichenbach, “Comparing landslide inventory maps,” Geomorphology, vol. 94, no. 3-4, pp. 268–289, 2008

    2008

  8. [8]

    High-resolution remote sensing image recognition of loess landslide: A case study of yan’an shaanxi province,

    D. Hui, Z. Sheng, Z. Hong, and Z. Tao, “High-resolution remote sensing image recognition of loess landslide: A case study of yan’an shaanxi province,” Northwest Geol., vol. 52, no. 3, pp. 231–239, 2019

    work page 2019

Show all 49 references
  1. [9]

    Landslide detection based on dem matching,

    R. Jie, J. Fei, X. Hua, W. Chao, and Z. Hong, “Landslide detection based on dem matching,” J. Surveying Mapping Sci. Technol., vol. 35, no. 5, pp. 477–484, 2018

    2018

  2. [10]

    Sentinel-1 sar amplitude imagery for rapid landslide detection,

    A. C. Mondini, M. Santangelo, M. Rocchetti, E. Ros- setto, A. Manconi, and O. Monserrat, “Sentinel-1 sar amplitude imagery for rapid landslide detection,” Remote sensing, vol. 11, no. 7, p. 760, 2019

    2019

  3. [11]

    A novel algorithm for calculating transition potential in cellular automata models of land-use/cover change,

    M. S. Roodposhti, J. Aryal, and B. A. Bryan, “A novel algorithm for calculating transition potential in cellular automata models of land-use/cover change,”Environmen- tal modelling & software , vol. 112, pp. 70–81, 2019

    2019

  4. [12]

    A comparative analysis of pixel- and object-based detection of landslides from very high- resolution images,

    R. N. Keyport, T. Oommen, T. R. Martha, K. S. Sajinku- mar, and J. S. Gierke, “A comparative analysis of pixel- and object-based detection of landslides from very high- resolution images,” International journal of applied earth observation and geoinformation , vol. 64, pp. 1–11, 2018

    2018

  5. [13]

    Postearthquake landslides mapping from landsat-8 data for the 2015 nepal earthquake using a pixel-based change detection method,

    W. Zhao, A. Li, X. Nan, Z. Zhang, and G. Lei, “Postearthquake landslides mapping from landsat-8 data for the 2015 nepal earthquake using a pixel-based change detection method,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 10, no. 5, p...

    2015

  6. [14]

    Landslide mapping from aerial photographs using change detection-based markov random field,

    Z. Li, W. Shi, P. Lu, L. Yan, Q. Wang, and Z. Miao, “Landslide mapping from aerial photographs using change detection-based markov random field,” Remote sensing of environment , vol. 187, pp. 76–90, 2016

    2016

  7. [15]

    Object-oriented mapping of landslides using random forests,

    A. Stumpf and N. Kerle, “Object-oriented mapping of landslides using random forests,” Remote sensing of environment, vol. 115, no. 10, pp. 2564–2577, 2011

    2011

  8. [16]

    Mapping land- slides on eo data: Performance of deep learning models vs. traditional machine learning models,

    N. Prakash, A. Manconi, and S. Loew, “Mapping land- slides on eo data: Performance of deep learning models vs. traditional machine learning models,” Remote Sens- ing, vol. 12, no. 3, p. 346, 2020

    2020

  9. [17]

    Landslide detection using residual networks and the fusion of spectral and topo- graphic information,

    M. I. Sameen and B. Pradhan, “Landslide detection using residual networks and the fusion of spectral and topo- graphic information,” Ieee Access , vol. 7, pp. 114 363– 114 373, 2019

    2019

  10. [18]

    Object based image analysis for remote sensing,

    T. Blaschke, “Object based image analysis for remote sensing,” ISPRS journal of photogrammetry and remote sensing, vol. 65, no. 1, pp. 2–16, 2010

    2010

  11. [19]

    Characterising spectral, spatial and morphometric properties of landslides for semi-automatic detection using object-oriented methods,

    T. R. Martha, N. Kerle, V . Jetten, C. J. van Westen, and K. V . Kumar, “Characterising spectral, spatial and morphometric properties of landslides for semi-automatic detection using object-oriented methods,” Geomorphol- ogy, vol. 116, no. 1-2, pp. 24–36, 2010

    2010

  12. [20]

    Landslide map- ping with multi-scale object-based image analysis–a case study in the baichi watershed, taiwan,

    T. Lahousse, K. Chang, and Y . Lin, “Landslide map- ping with multi-scale object-based image analysis–a case study in the baichi watershed, taiwan,” Natural Hazards and Earth System Sciences , vol. 11, no. 10, pp. 2715– 2726, 2011

    2011

  13. [21]

    A semi-automated object-based approach for landslide detection validated by persistent scatterer interferometry measures and landslide inventories,

    D. H ¨olbling, P. F¨ureder, F. Antolini, F. Cigna, N. Casagli, and S. Lang, “A semi-automated object-based approach for landslide detection validated by persistent scatterer interferometry measures and landslide inventories,” Re- mote Sensing , vol. 4, no. 5, pp. 1310–1336, 2012

    2012

  14. [22]

    Multi-aspect analysis of object-oriented landslide detec- tion based on an extended set of lidar-derived terrain fea- tures,

    K. Pawłuszek, S. Marczak, A. Borkowski, and P. Tarolli, “Multi-aspect analysis of object-oriented landslide detec- tion based on an extended set of lidar-derived terrain fea- tures,” ISPRS international journal of geo-information , vol. 8, no. 8, p. 321, 2019

    2019

  15. [23]

    Mask r-cnn,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask r-cnn,” in Proceedings of the IEEE international confer- ence on computer vision , 2017, pp. 2961–2969

    2017

  16. [24]

    Fast r-cnn,

    R. Girshick, “Fast r-cnn,” in Proceedings of the IEEE international conference on computer vision , 2015, pp. 1440–1448

    2015

  17. [25]

    Faster r-cnn: Towards real-time object detection with region proposal networks,

    S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks,” Advances in neural information processing systems, vol. 28, 2015

    2015

  18. [26]

    Fully convolu- tional networks for semantic segmentation,

    J. Long, E. Shelhamer, and T. Darrell, “Fully convolu- tional networks for semantic segmentation,” in Proceed- ings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 3431–3440

    2015

  19. [27]

    Encoder-decoder with atrous separable convo- lution for semantic image segmentation,

    L.-C. Chen, Y . Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder-decoder with atrous separable convo- lution for semantic image segmentation,” in Proceedings of the European conference on computer vision (ECCV) , 2018, pp. 801–818

    2018

  20. [28]

    Gan- based siamese framework for landslide inventory map- ping using bi-temporal optical remote sensing images,

    B. Fang, G. Chen, L. Pan, R. Kou, and L. Wang, “Gan- based siamese framework for landslide inventory map- ping using bi-temporal optical remote sensing images,” IEEE Geoscience and Remote Sensing Letters , vol. 18, no. 3, pp. 391–395, 2020

    2020

  21. [29]

    Landslide segmentation with u-net: Evaluating differ- ent sampling methods and patch sizes,

    L. P. Soares, H. C. Dias, and C. H. Grohmann, “Landslide segmentation with u-net: Evaluating differ- ent sampling methods and patch sizes,” arXiv preprint arXiv:2007.06672, 2020

    2007 arXiv

  22. [30]

    Landslide de- tection using densely connected convolutional networks and environmental conditions,

    H. Cai, T. Chen, R. Niu, and A. Plaza, “Landslide de- tection using densely connected convolutional networks and environmental conditions,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sens- ing, vol. 14, pp. 5235–5247, 2021

    2021

  23. [31]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weis- senborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Min- derer, G. Heigold, S. Gelly et al. , “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929 , 2020

    2010 arXiv

  24. [32]

    Swin transformer: Hierarchical vi- sion transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12 S. Lin, and B. Guo, “Swin transformer: Hierarchical vi- sion transformer using shifted windows,” in Proceedings of the IEEE/CVF international conference on computer...

    2021

  25. [33]

    Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,

    S. Zheng, J. Lu, H. Zhao, X. Zhu, Z. Luo, Y . Wang, Y . Fu, J. Feng, T. Xiang, P. H. Torr et al., “Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition ...

    2021

  26. [34]

    Seg- menter: Transformer for semantic segmentation,

    R. Strudel, R. Garcia, I. Laptev, and C. Schmid, “Seg- menter: Transformer for semantic segmentation,” in Pro- ceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 7262–7272

    2021

  27. [35]

    Segformer: Simple and efficient design for semantic segmentation with transformers,

    E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “Segformer: Simple and efficient design for semantic segmentation with transformers,” Advances in neural information processing systems , vol. 34, pp. 12 077–12 090, 2021

    2021

  28. [36]

    Landslide susceptibility prediction based on im- age semantic segmentation,

    B. Du, Z. Zhao, X. Hu, G. Wu, L. Han, L. Sun, and Q. Gao, “Landslide susceptibility prediction based on im- age semantic segmentation,” Computers & Geosciences , vol. 155, p. 104860, 2021

    2021

  29. [37]

    Research progress and prospect on reactivation of ancient landslides,

    Z. Yongshuang, W. Ruian, G. Changbao, W. Lichao, Y . Xin, and Y . Zhihua, “Research progress and prospect on reactivation of ancient landslides,” Advances in Earth Science, vol. 33, no. 7, p. 728, 2018

    2018

  30. [38]

    Simmim: A simple framework for masked image modeling,

    Z. Xie, Z. Zhang, Y . Cao, Y . Lin, J. Bao, Z. Yao, Q. Dai, and H. Hu, “Simmim: A simple framework for masked image modeling,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 9653–9663

    2022

  31. [39]

    Unsupervised feature learning via non-parametric instance discrimina- tion,

    Z. Wu, Y . Xiong, S. X. Yu, and D. Lin, “Unsupervised feature learning via non-parametric instance discrimina- tion,” in Proceedings of the IEEE conference on com- puter vision and pattern recognition , 2018, pp. 3733– 3742

    2018

  32. [40]

    An iterative classification and semantic seg- mentation network for old landslide detection using high- resolution remote sensing images,

    Z. Lu, Y . Peng, W. Li, J. Yu, D. Ge, L. Han, and W. Xiang, “An iterative classification and semantic seg- mentation network for old landslide detection using high- resolution remote sensing images,” IEEE Transactions on Geoscience and Remote Sensing , 2023

    2023

  33. [41]

    Bootstrap your own latent- a new approach to self-supervised learning,

    J.-B. Grill, F. Strub, F. Altch ´e, C. Tallec, P. Richemond, E. Buchatskaya, C. Doersch, B. Avila Pires, Z. Guo, M. Gheshlaghi Azar et al. , “Bootstrap your own latent- a new approach to self-supervised learning,” Advances in neural information processing systems , vol. 33, pp...

    2020

  34. [42]

    Emerging properties in self-supervised vision transformers,

    M. Caron, H. Touvron, I. Misra, H. J ´egou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 9650–9660

    2021

  35. [43]

    Self-attention with relative position representations,

    P. Shaw, J. Uszkoreit, and A. Vaswani, “Self-attention with relative position representations,” arXiv preprint arXiv:1803.02155, 2018

    2018 arXiv

  36. [44]

    Squeeze-and-excitation networks,

    J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 7132– 7141

    2018

  37. [45]

    Spatial transformer networks,

    M. Jaderberg, K. Simonyan, A. Zisserman et al., “Spatial transformer networks,” Advances in neural information processing systems, vol. 28, 2015

    2015

  38. [46]

    Bisdenet: A new lightweight deep learning-based framework for efficient landslide detection,

    T. Chen, X. Gao, G. Liu, C. Wang, Z. Zhao, J. Dou, R. Niu, and A. Plaza, “Bisdenet: A new lightweight deep learning-based framework for efficient landslide detection,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2024

    2024

  39. [47]

    A new mask r-cnn-based method for improved landslide detection,

    S. L. Ullo, A. Mohan, A. Sebastianelli, S. E. Ahamed, B. Kumar, R. Dwivedi, and G. R. Sinha, “A new mask r-cnn-based method for improved landslide detection,” IEEE Journal of Selected Topics in Applied Earth Ob- servations and Remote Sensing , vol. 14, pp. 3799–3810, 2021

    2021

  40. [48]

    Tcnet: Multiscale fusion of transformer and cnn for semantic segmentation of remote sensing images,

    X. Xiang, W. Gong, S. Li, J. Chen, and T. Ren, “Tcnet: Multiscale fusion of transformer and cnn for semantic segmentation of remote sensing images,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2024

    2024

  41. [49]

    Grad-cam: Visual explanations from deep networks via gradient-based localization,

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-cam: Visual explanations from deep networks via gradient-based localization,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 618–626

    2017

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.