Pith. sign in

REVIEW 3 major objections 5 minor 84 references

Segmenting Low-Contrast XCTs of Concrete: An Unsupervised Approach

T0 review · 3 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read This paper shows that a U-Net trained with self-annotated superpixel labels can separate aggregate from mortar in low-contrast concrete XCT images, outperforming direct greyscale thresholding on out-of-distribution data.

desk verdict Useful, honest adaptation of Kanezaki-style self-annotation to low-contrast concrete XCT, but the headline claim against direct thresholding is not yet controlled because of manual per-image thresholds and asymmetric postprocessing. read the letter →

arxiv 2603.00127 v2 pith:HSALGSXG submitted 2026-02-23 cs.CV

Kaustav Das , Gaston Rauchs , Jan Sykora , Anna Kucerova This is my paper · ORCID
classification cs.CV
keywords X-raycomputedtomographyconcreteunsupervisedsemanticsegmentationself-annotationsuperpixelsU-Netaggregatelow-contrastXCT
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 tries to establish that low-contrast X-ray CT images of concrete—where aggregate and mortar look almost identical in intensity—can be segmented without any labelled training data, by having the model teach itself. The proposed loop uses a convolutional encoder-decoder network to produce rough per-pixel phase predictions, then SLIC superpixels to smooth those predictions into region labels by majority vote, and then trains the network on those self-generated labels as if they were ground truth. On a held-out concrete sample, the resulting aggregate maps beat direct greyscale thresholding on the metrics reported. This matters because concrete XCT segmentation normally requires costly manual annotation or contrast-enhancing additives, both of which limit practical use.

What carries the argument

The engine is the superpixel-refinement self-annotation loop. SLIC partitions each image into small contiguous regions of perceptual similarity; the most frequent predicted class in each region becomes the label for all its pixels, producing dynamic image labels that are fed back as training targets. This converts noisy, pixel-level CNN outputs into spatially coherent pseudo-labels and ties local superpixel structure to the global context captured by the network's receptive field. Per-channel normalization of the output prevents the model from collapsing to a single phase, although the paper notes it remains fragile under class imbalance. The architecture is a U-Net-style encoder-decoder wit

What would settle it

Compare the superpixel-refined training labels against the 16 manually annotated slices: if superpixels frequently straddle aggregate-mortar boundaries in low-contrast regions and the model's errors correlate with those straddles, the refinement assumption fails. More directly, train the same pipeline with deliberately misaligned superpixels (e.g., extreme compactness or oversized superpixels); if aggregate identification accuracy does not degrade, the superpixel-alignment assumption is not load-bearing.

Watch

Extended reading notes

Core claim

The paper's central claim is that a convolutional encoder-decoder network can learn to separate aggregate from mortar in low-contrast concrete XCT images using no labelled data for those phases, and that the resulting aggregate segmentations are more accurate than direct greyscale thresholding. The self-annotation loop works as follows: after each training pass, per-pixel predictions are normalized per output channel, converted to hard labels by argmax, and then refined by replacing every pixel's label with the majority label of its SLIC superpixel. These refined labels—called dynamic image labels—are used as the cross-entropy training targets for the next iteration, evolving as the model im

Load-bearing premise

The load-bearing premise is that SLIC superpixels stay inside single phases and that the majority predicted class within a superpixel is the true phase; in low-contrast regions where superpixels straddle aggregate-mortar boundaries, wrong labels are fed back into training and the model can learn to reproduce those errors.

Editorial extensions

If this is right

  • Concrete XCT scans can be segmented into aggregate and mortar phases without manual annotation, removing the main bottleneck to using XCT in concrete mechanics.
  • Thresholding the aggregate-specific channel of the normalized network output identifies aggregates better than thresholding raw XCT slices, so the network acts as an intensity transform that sharpens phase separability.
  • A purely unsupervised three-phase setup fails to isolate porosity; the semi-supervised variant, which thresholds only the easily separable pore phase, gives unambiguous three-phase labels.
  • The learned segmentation can bootstrap supervised training: the paper explicitly proposes using its outputs as initial labels for fine-tuning on new concrete datasets.
  • The method inherits concrete failure modes: aggregates near the sample periphery are missed and closely spaced small aggregates are merged, so downstream users should expect those behaviors.

Reading between the lines

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

  • A natural testable extension is to apply the same self-annotation loop to other low-contrast multiphase materials with a known phase count and check whether the superpixel-mode assumption degrades as phase-boundary contrast drops.
  • The reported periphery failure may be a receptive-field or padding artifact rather than a fundamental limit; training on full slices or using border-aware crops could recover those aggregates.
  • Because the paper flags per-channel normalization as fragile under class imbalance, a class-balanced or entropy-regularized alternative could make the fully unsupervised three-phase case succeed and remove the need to threshold porosity.
  • The dynamic-label loop is a form of self-training, so its convergence likely depends on initial random predictions being slightly better than chance within each superpixel; this could be tested by seeding training with a simple intensity prior.
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

3 major / 5 minor

Summary. The paper presents a self-annotation-based segmentation pipeline for low-contrast X-ray CT images of concrete, targeting three phases: aggregate, mortar, and porosity. A U-Net is trained without manually labeled aggregate/mortar data; per-pixel predictions are converted into 'dynamic image labels' by per-channel normalization, argmax classification, and SLIC superpixel refinement (Eq. 12, Section 3.4). Three training configurations are tested: fully unsupervised with 3 channels (US3), fully unsupervised with 4 channels (US4), and semi-supervised with 3 channels where porosity labels are provided by thresholding (SS3). On a held-out XCT sample with 16 manually annotated slices, the authors compare aggregate segmentation obtained by thresholding the aggregate-specific logit channel of SS3 against direct grayscale thresholding of the original XCT images, reporting that the proposed method consistently outperforms direct thresholding across IoU, precision, sensitivity, and F1. The authors also document limitations, including failure of the fully unsupervised variants to resolve porosity and systematic errors near sample peripheries and between closely spaced aggregates.

Significance. If the empirical claim is substantiated, the work has practical value: it offers a way to train a segmentation model for concrete XCT without manually annotated aggregate/mortar labels, which is a real bottleneck for morphology-based analysis in experimental mechanics. The paper also ships processed data, code, and model weights via Zenodo, which supports reproducibility and follow-up work. The held-out manual ground truth provides a meaningful independent check on the self-annotation loop. However, the headline claim is currently under-supported because the comparison against direct thresholding is not controlled, and the only fully successful configuration (SS3) is semi-supervised rather than unsupervised. The contribution of the self-annotation mechanism itself, as distinct from per-image threshold tuning and heavier morphological postprocessing, is not yet isolated.

major comments (3)
  1. [§7.2, Fig. 31; §6; §5.3] The central claim that logit thresholding of the SS3 output 'consistently outperformed' direct grayscale thresholding is confounded. Section 6 states that the threshold applied to the aggregate channel is 'unique' for each image but gives no selection rule, and Section 5.3 explicitly omits thresholding details. Direct thresholding also appears to be manually tuned. Furthermore, the postprocessing chains differ: the logit pipeline applies hole filling to 512 px, binary erosion, removal of objects <128 px, and final dilation, while the direct-threshold baseline only removes objects <128 px and holes <256 px. Reported gains could therefore arise from per-image threshold selection and extra morphological operations rather than from the learned aggregate/mortar separation. Please fix both operations to a reproducible rule (e.g., Otsu or a fixed percentile on each channel) and use identical po
  2. [§5.1–§5.3; Abstract] The paper is framed as an 'unsupervised approach,' but the fully unsupervised variants US3 and US4 both fail to unambiguously segment the porous phase (Figs. 21–24). The successful SS3 configuration uses thresholded porosity labels during training, making it semi-supervised. The abstract's claim that the methodology consistently outperforms direct thresholding refers to SS3, not to the unsupervised variants. This is a load-bearing mismatch between the title/abstract and the experimental results. Either temper the 'unsupervised' claims to 'semi-supervised for porosity, unsupervised for aggregate/mortar,' or demonstrate a fully unsupervised configuration that resolves all three phases.
  3. [§3.4, Eq. (12)] The self-annotation mechanism rests on the assumption that the most frequent predicted class within a SLIC superpixel is the correct semantic label for all pixels in that superpixel. If superpixels cross aggregate/mortar boundaries—which is likely in low-contrast regions—incorrect labels are fed back as training targets. The assumption is stated but never validated, e.g., by comparing superpixel boundaries against the manually annotated test slices or by measuring boundary recall/precision. A quantitative sanity check on the 16 annotated slices, or a sensitivity analysis over SLIC compactness and superpixel count, would materially strengthen the claim that the self-annotation loop is learning meaningful phase boundaries rather than propagating its own errors.
minor comments (5)
  1. [§2.1] Typo: 'numcerical' should be 'numerical.' Also, the description of beam-hardening correction would benefit from a reference or more detail on how the manual parameter choice affects downstream slices.
  2. [§5, Fig. 20] The loss curves are averaged over 10 training runs, but the qualitative results are shown for a single unspecified run. Please state clearly whether the presented segmentation examples are representative of all runs and whether the quantitative evaluation in Fig. 31 uses one model or an ensemble/average.
  3. [§7.2] Only aggregate-phase metrics are reported. While mortar is complementary to aggregate, porosity is not evaluated quantitatively. Since porosity is a key phase in concrete morphology and SS3 is the only configuration that resolves it, reporting porosity IoU on the same ground truth would make the evaluation more complete.
  4. [Eq. (10)] The normalization in Eq. (10) uses the per-image mean and variance of each channel, but the notation does not make clear whether these statistics are computed per tile, per image, or per batch. Please clarify, as this affects the interpretation of the 'dynamic labels' in minibatch training.
  5. [§6] The suggestion to interpolate thresholds across slices using PCHIP is useful, but the manuscript does not state how many slices are needed for reliable interpolation or how sensitive the final metrics are to the choice of interpolation points. A short experiment or guidance would improve practical reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central empirical claim is grounded in held-out manual annotations, and the self-annotation loop is an explicit training mechanism, not a disguised restatement of the result.

full rationale

The self-annotation procedure (Eqs. 10-12) does define training targets as a function of the model's own normalized predictions (argmax followed by superpixel mode), so in isolation the training objective is self-referential. However, the paper's central claim is not that these dynamic labels are correct by construction; it is that a U-Net trained with this objective generalizes to a held-out XCT sample (TPR) and, after postprocessing, matches manually annotated ground truth better than direct greyscale thresholding. That evaluation uses external manual annotations, so the comparison does not reduce to the model's own labels. The postprocessing threshold is described as 'unique' per image with no selection rule, and the morphological postprocessing differs between the proposed and baseline pipelines; this is a reproducibility/confounding concern about the empirical comparison, not a definitional circularity. No load-bearing self-citations or imported uniqueness/ansatz claims were found. The weaknesses acknowledged in Section 8 (porosity ambiguity, periphery aggregates, small-aggregate clumping, per-channel normalization) are honest limitations rather than circular reasoning.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

No new physical entities are introduced. The main free parameters are algorithm hyperparameters (SLIC compactness, number of superpixels) and postprocessing thresholds, several of which are chosen manually and could affect the reported performance. The strongest assumption is that superpixel majority-vote labels are correct, which underpins the whole self-annotation loop.

free parameters (5)
  • SLIC compactness
    Chosen through experimental trials to make superpixel boundaries adhere to phase boundaries (Section 4.1).
  • Number of superpixels
    Set so superpixels are ~4mm equivalent, based on the convention that aggregates smaller than 4mm are part of mortar (Section 4.1).
  • Threshold for halo removal
    Manually chosen per sample to set non-solid phase voxels to zero (Section 2.2).
  • Threshold for aggregate channel binarization
    Manual per-image threshold used in postprocessing to generate the final aggregate mask; selection rule not specified (Section 6).
  • Morphological postprocessing parameters = fill holes ≤512px, remove objects <128px, final dilation, remove <256px
    Hand-set parameters to denoise the binary aggregate mask (Section 7).
assumptions (4)
  • domain assumption SLIC superpixels correspond to perceptually similar regions that align with phase boundaries
    Assumed in Section 3.4; if superpixels cross aggregate/mortar boundaries, the dynamic labels are noisy.
  • ad hoc to paper The most frequent class within a superpixel is the correct semantic label for all pixels in it
    Core self-annotation rule in Section 3.4; a strong assumption that is not validated against ground truth during training.
  • domain assumption Concrete can be segmented into three phases: aggregate, mortar, and porosity
    Standard in concrete mesostructure analysis, stated in Section 1.
  • domain assumption The receptive field of the U-Net is sufficient to capture global context for each pixel
    Implied in Section 3.2; the receptive field is used to relate local superpixels to global context, but no analysis is provided to confirm the field size is adequate for concrete features.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Segmenting Low-Contrast XCTs of Concrete: An Unsupervised Approach." pith.science (2026). https://pith.science/paper/HSALGSXG

@misc{pith2026260300127,
  author       = {Pith},
  title        = {Pith review of: Segmenting Low-Contrast XCTs of Concrete: An Unsupervised Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HSALGSXG}},
  note         = {Machine review of arXiv:2603.00127}
}
read the original abstract

X-Ray Computed Tomography (XCT) is a compelling tool in experimental mechanics, capable of non-destructively extracting information pertaining to the internal morphology of materials. For materials with random heterogeneous morphology such as concrete, such information is of particular relevance since it allows for studies of morphology-related behaviour and for predictive modelling. Nevertheless, XCT images require semantic segmentation for practical usage. Here, concrete poses a unique challenge due to the similar X-ray attenuation coefficients of aggregates and mortar, which result in low contrast between the two phases in the ensuing XCT images. As such, purely intensity-dependent semantic segmentation tools remain unfeasible. While vision transformers (ViTs) and convolutional neural networks (CNNs) are proven techniques for semantic segmentation in such challenging cases, they typically require labelled training data, which is often unavailable for concrete or resource-intensive to obtain, thereby limiting their relevance. To address this challenge, a self-annotation technique is presented here that leverages superpixel algorithms to identify perceptually similar local regions in an image and relates them to the global context by utilizing the receptive field of a CNN-based model. This enables the model to learn a global-local relationship in the images and facilitates the identification of semantically similar structures. When evaluated against manually annotated ground truth on out-of-distribution data, the proposed methodology consistently outperformed direct greyscale thresholding across all pertinent metrics, demonstrating improved discernibility between aggregates and mortar, and providing the most favourable balance of sensitivity and precision for aggregate-phase identification.

Figures

Figures reproduced from arXiv: 2603.00127 by the authors.

Figure 1
Figure 1. Example of the concrete XCT specimen in 3D. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Example of low-contrast regions in an XCT image of concrete. (a)–(c) illustrate aggregates with only part of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Inter-sample variation of voxel intensity depicted using histograms. For each sample, the peak to the left [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (28 more)
Figure 4
Figure 4. Figure 4: Illustration of variations in pixel values typically observed in all samples shown using selected slices from one [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Examples of concentric ring artifact (a), and bluring (b) seen near the centre of the cylinder and halo-effect (c), [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Maximum pixel intensity measured along the magenta line (left image) for the entire sample (along the cylinder [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Example of an original image (a) and its preprocessed version (b). Notably, the halo-like artifacts visible in the [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: The U-Net model architecture. Green rectangles correspond to feature tensors. Input (grey rectangle) is the [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: Convolution block (a) and Upsampling block (b). Padding size, kernel size, stride and dropout probability are [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 10
Figure 10. Figure 10: A single-channel input tensor undergoes convolution using kernels of size (3,3) and a stride of (1,1) in (A). [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 11
Figure 11. Figure 11: 2D transposed convolution on a single channel input, i.e. [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]
Figure 12
Figure 12. Figure 12: Illustration of a receptive field. The highlighted pixel in (d) is the result of applying a convolution with a (3 [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 13
Figure 13. Figure 13: Example of arg max application on the prediction values of three randomly selected training samples in the first training epoch. Each colour is representative of the categorical class with the maximum prediction value. This necessitates identification of localized reg…
Figure 14
Figure 14. Figure 14: Comparison of superpixels obtained using [PITH_FULL_IMAGE:figures/full_fig_p014_14.png]
Figure 15
Figure 15. Figure 15: A schematic depiction of the superpixel refinement operation. Given an image with categorical pixels as in (a), [PITH_FULL_IMAGE:figures/full_fig_p014_15.png]
Figure 16
Figure 16. Figure 16: Superpixel refinement as part of the overall scheme. Given an image, the superpixel map is computed and for [PITH_FULL_IMAGE:figures/full_fig_p015_16.png]
Figure 17
Figure 17. Figure 17: Illustration of evolution of dynamic image labels with respect to training iterations (Note: Number of images do [PITH_FULL_IMAGE:figures/full_fig_p015_17.png]
Figure 18
Figure 18. Figure 18: The alternative approach of rescaling the images to a smaller dimension instead of tiling was [PITH_FULL_IMAGE:figures/full_fig_p016_18.png]
Figure 19
Figure 19. Figure 19: Schematic flowchart for unsupervised training of the U-Net model aided by dynamic label generation assuming [PITH_FULL_IMAGE:figures/full_fig_p017_19.png]
Figure 20
Figure 20. Figure 20: Mean training losses and mean validation losses averaged over [PITH_FULL_IMAGE:figures/full_fig_p019_20.png]
Figure 21
Figure 21. Figure 21: US3: Example of 3-channel prediction results using model trained in an unsupervised manner for the given input slices (left). Column (a) and (c) can be associated with the mortar and aggregate phases respectively. Channel (b) is ambiguous and displays similar scores f…
Figure 22
Figure 22. Figure 22: US3: arg max classification of Y¯ PR US3 and the superpixel refinement L˜ PR US3 using parameters equivalent to that used during training. Slice 125 Slice 250 [PITH_FULL_IMAGE:figures/full_fig_p020_22.png]
Figure 23
Figure 23. Figure 23: US4: Example of 4-channel prediction results using model trained in the unsupervised scenario for input slices presented in [PITH_FULL_IMAGE:figures/full_fig_p020_23.png]
Figure 24
Figure 24. Figure 24: US4: arg max classification of Y¯ PR US4 and the superpixel refinement L˜ PR US4 using parameters equivalent to that used during training. 5.2. Unsupervised Training: Given 4-phases (US4) The constraint on the number of output features was relaxed to 4 to investigate …
Figure 25
Figure 25. Figure 25: SS3: Example of 3-channel prediction results using the model trained in the semi-supervised manner for the given input slices (left). Column (a), (b) and (c) can unambiguously be associated with the aggregate, mortar and porosity respectively. Slice 125 Slice 250 Argm…
Figure 26
Figure 26. Figure 26: SS3: arg max classification of Y¯ PR SS3 and the superpixel refinement using parameters equivalent to that used during training. superpixel refinements in this case is not identical to the tensor L˜ t in Section 3 due to difference in the annotation procedure. Normali…
Figure 27
Figure 27. Figure 27: Original XCT slices from TPR (a). Example of 3-phase segmentation (b) using a thresholding-based postpro￾cessing approach. Aggregates identified in white, mortar in gray and porosity in black. Comparison of Aggregates obtained following the process against ground trut…
Figure 28
Figure 28. Figure 28: Comparison of binarized results of aggregates. (a) Input slices, (b) Ground Truth, (c) Aggregates obtained [PITH_FULL_IMAGE:figures/full_fig_p024_28.png]
Figure 29
Figure 29. Figure 29: (a) Model fails to identify aggregates at the periphery of the cylindrical casting. (b) In case of aggregates located [PITH_FULL_IMAGE:figures/full_fig_p024_29.png]
Figure 30
Figure 30. Figure 30: Interpretation of colours and definition of performance metrics. [PITH_FULL_IMAGE:figures/full_fig_p025_30.png]
Figure 31
Figure 31. Figure 31: Quantitative evaluation of segmentation quality focusing on the aggregates compared to the ground truth of [PITH_FULL_IMAGE:figures/full_fig_p026_31.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

84 extracted references · 4 linked inside Pith

  1. [1]

    Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., and Süsstrunk, S. (2010). Slic superpixels

    2010

  2. [2]

    Bajcsy, P., Feldman, S., Majurski, M., Snyder, K., and Brady, M. (2020). Approaches to training multiclass semantic image segmentation of damage in concrete.Journal of Microscopy, 279(2):98–113

    2020

  3. [3]

    S., Wang, C., Zhou, X., and Hassan, M

    Bangaru, S. S., Wang, C., Zhou, X., and Hassan, M. (2022). Scanning electron microscopy (sem) image segmentation for microstructure analysis of concrete using u-net convolutional neural network. Automation in Construction, 144:104602

    2022

  4. [4]

    Bengio, Y. (2012). Practical recommendations for gradient-based training of deep architectures. In Neural networks: Tricks of the trade: Second edition, pages 437–478. Springer

    2012

  5. [5]

    Bishop, C. M. and Bishop, H. (2023).Deep learning: Foundations and concepts. Springer Nature

    2023

  6. [6]

    and van Mier, J

    Bisschop, J. and van Mier, J. G. (2008). Effect of aggregates and microcracks on the drying rate of cementitious composites.Cement and Concrete Research, 38(10):1190–1196

    2008

  7. [7]

    Böhm, R., Stiller, J., Behnisch, T., Zscheyge, M., Protz, R., Radloff, S., Gude, M., and Hufenbach, W. (2015). A quantitative comparison of the capabilities of in situ computed tomography and conventional computed tomography for damage analysis of composites.Composites Science and Technology, 110:62– 68

    2015

  8. [8]

    C., Apley, D

    Bostanabad, R., Zhang, Y., Li, X., Kearney, T., Brinson, L. C., Apley, D. W., Liu, W. K., and Chen, W. (2018). Computational microstructure characterization and reconstruction: Review of the state-of-the-art techniques.Progress in Materials Science, 95:1–41

    2018

Show all 84 references
  1. [9]

    Bottou, L. (2012). Stochastic gradient descent tricks. InNeural networks: tricks of the trade: second edition, pages 421–436. Springer

    2012

  2. [10]

    Brisard, S., Serdar, M., and Monteiro, P. J. (2020). Multiscale x-ray tomography of cementitious materials: A review.Cement and Concrete Research, 128:105824

    2020

  3. [11]

    P., Lunardelli, M., Leusmann, T., Varady, P

    Carrara, P., Kruse, R., Bentz, D. P., Lunardelli, M., Leusmann, T., Varady, P. A., and De Lorenzis, L. (2018). Improved mesoscale segmentation of concrete from 3d x-ray images using contrast enhancers. Cement and Concrete Composites, 93:30–42

    2018

  4. [12]

    Chen, B., Zhang, H., Wang, G., Huo, J., Li, Y., and Li, L. (2023). Automatic concrete infrastructure crack semantic segmentation using deep learning.Automation in Construction, 152:104950

    2023

  5. [13]

    Chen, L., Shan, W., and Liu, P. (2021). Identification of concrete aggregates using k-means clustering and level set method. InStructures, volume 34, pages 2069–2076. Elsevier

    2021

  6. [14]

    Cheng, B., Schwing, A., and Kirillov, A. (2021). Per-pixel classification is not all you need for semantic segmentation.Advances in neural information processing systems, 34:17864–17875

    2021

  7. [15]

    Coenen, M., Schack, T., Beyer, D., Heipke, C., and Haist, M. (2021). Semi-supervised segmenta- tion of concrete aggregate using consensus regularisation and prior guidance.ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences V-2-2021 (2021), 2:83–91

    2021

  8. [16]

    L., Nardelli, M

    Curtarolo, S., Hart, G. L., Nardelli, M. B., Mingo, N., Sanvito, S., and Levy, O. (2013). The high-throughput highway to computational materials design.Nature materials, 12(3):191–201

    2013

  9. [17]

    Dong, Y., Su, C., Qiao, P., and Sun, L. (2020). Microstructural crack segmentation of three- dimensional concrete images based on deep convolutional neural networks.Construction and Building Materials, 253:119185

    2020

  10. [18]

    Fritsch, F. N. and Butland, J. (1984). A method for constructing local monotone piecewise cubic interpolants.SIAM journal on scientific and statistical computing, 5(2):300–304. 28

    1984

  11. [19]

    Fujiwara, T. (2008). Effect of aggregate on drying shrinkage of concrete.Journal of Advanced Concrete Technology, 6(1):31–44

    2008

  12. [20]

    T., Niezgoda, S

    Fullwood, D. T., Niezgoda, S. R., Adams, B. L., and Kalidindi, S. R. (2010). Microstructure sensitive design for performance optimization.Progress in Materials Science, 55(6):477–562

    2010

  13. [21]

    and Bengio, Y

    Glorot, X. and Bengio, Y. (2010). Understanding the difficulty of training deep feedforward neu- ral networks. InProceedings of the thirteenth international conference on artificial intelligence and statistics, pages 249–256. JMLR Workshop and Conference Proceedings

    2010

  14. [22]

    Gonzalez, R. C. (2009).Digital image processing. Pearson education india

    2009

  15. [23]

    (2016).Deep Learning

    Goodfellow, I., Bengio, Y., and Courville, A. (2016).Deep Learning. MIT Press.http://www. deeplearningbook.org

    2016

  16. [24]

    and Jirásek, M

    Havlásek, P. and Jirásek, M. (2016). Multiscale modeling of drying shrinkage and creep of concrete. Cement and concrete research, 85:55–74

    2016

  17. [25]

    Havlásek, P., Šmilauer, V., Dohnalová, L., and Sovják, R. (2021). Shrinkage-induced deformations and creep of structural concrete: 1-year measurements and numerical prediction.Cement and Concrete Research, 144:106402

    2021

  18. [26]

    He, K., Gkioxari, G., Dollár, P., and Girshick, R. (2017). Mask r-cnn. InProceedings of the IEEE international conference on computer vision, pages 2961–2969

    2017

  19. [27]

    Hilloulin, B., Bekrine, I., Schmitt, E., and Loukili, A. (2022). Modular deep learning segmentation algorithm for concrete microscopic images.Construction and Building Materials, 349:128736

    2022

  20. [28]

    Hufenbach, W., Böhm, R., Gude, M., Berthel, M., Hornig, A., Ručevskis, S., and Andrich, M. (2012). A test device for damage characterisation of composites based on in situ computed tomography. Composites Science and Technology, 72(12):1361–1367

    2012

  21. [29]

    Idiart, A., Bisschop, J., Caballero, A., and Lura, P. (2012). A numerical and experimental study of aggregate-induced shrinkage cracking in cementitious composites.Cement and Concrete Research, 42(2):272–281

    2012

  22. [30]

    Batchnormalization: Acceleratingdeepnetworktrainingbyreducing internal covariate shift

    Ioffe, S.andSzegedy, C.(2015). Batchnormalization: Acceleratingdeepnetworktrainingbyreducing internal covariate shift. InInternational conference on machine learning, pages 448–456. pmlr

    2015

  23. [31]

    Jung, C., Müsebeck, F., Barisin, T., Schladitz, K., Redenbach, C., Kiesche, M., and Pahn, M. (2022). Towards automatic crack segmentation in 3d concrete images. In11th conf. ind. comput. tomogr.(ICT)

    2022

  24. [32]

    R., Buzzy, M., Boyce, B

    Kalidindi, S. R., Buzzy, M., Boyce, B. L., and Dingreville, R. (2022). Digital twins for materials. Frontiers in Materials, 9:818535

    2022

  25. [33]

    F., Simpson, J

    Kamnitsas, K., Ledig, C., Newcombe, V. F., Simpson, J. P., Kane, A. D., Menon, D. K., Rueckert, D., and Glocker, B. (2017). Efficient multi-scale 3d cnn with fully connected crf for accurate brain lesion segmentation.Medical image analysis, 36:61–78

    2017

  26. [34]

    Kanezaki, A. (2018). Unsupervised image segmentation by backpropagation. In2018 IEEE inter- national conference on acoustics, speech and signal processing (ICASSP), pages 1543–1547. IEEE

    2018

  27. [35]

    S., Mudigere, D., Nocedal, J., Smelyanskiy, M., and Tang, P

    Keskar, N. S., Mudigere, D., Nocedal, J., Smelyanskiy, M., and Tang, P. T. P. (2016). On large-batch training for deep learning: Generalization gap and sharp minima.arXiv preprint arXiv:1609.04836

    2016 arXiv

  28. [36]

    H., Park, K., et al

    Kim, H.-T., Han, T.-S., Lee, Y. H., Park, K., et al. (2024). Recursive aggregate segmentation by erosion and reconstitution (raser) to characterize concrete microstructure using complementarity of x-ray and neutron computed tomography.Cement and Concrete Composites, 148:105437. 29

    2024

  29. [37]

    A., Park, K., et al

    Kim, H.-T., Szilágyi, V., Kis, Z., Szentmiklósi, L., Glinicki, M. A., Park, K., et al. (2021). Recon- struction of concrete microstructure using complementarity of x-ray and neutron tomography.Cement and Concrete Research, 148:106540

    2021

  30. [38]

    Kim, W., Kanezaki, A., and Tanaka, M. (2020). Unsupervised learning of image segmentation based on differentiable feature clustering.IEEE Transactions on Image Processing, 29:8055–8068

    2020

  31. [39]

    C., Lo, W.-Y., et al

    Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W.-Y., et al. (2023). Segment anything. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 4015–4026

    2023

  32. [40]

    Kumar, P., Sharma, A., and Kota, S. R. (2021). Automatic multiclass instance segmentation of concrete damage using deep learning model.IEEE Access, 9:90330–90345

    2021

  33. [41]

    and Bhattacharjee, B

    Kumar, R. and Bhattacharjee, B. (2003). Porosity, pore size distribution and in situ strength of concrete.Cement and concrete research, 33(1):155–164

    2003

  34. [42]

    S., Henderson, D., Howard, R

    LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., and Jackel, L. D. (1989). Backpropagation applied to handwritten zip code recognition.Neural computation, 1(4):541–551

    1989

  35. [43]

    LeCun, Y., Bottou, L., Bengio, Y., and Haffner, P. (1998). Gradient-based learning applied to document recognition.Proceedings of the IEEE, 86(11):2278–2324

    1998

  36. [44]

    C., and Tang, X

    Liu, Z., Li, X., Luo, P., Loy, C. C., and Tang, X. (2017). Deep learning markov random field for semantic segmentation.IEEE transactions on pattern analysis and machine intelligence, 40(8):1814– 1828

    2017

  37. [45]

    M., Qiu, Y., Martin, B., Heard, W., Williams, B., and Nie, X

    Loeffler, C. M., Qiu, Y., Martin, B., Heard, W., Williams, B., and Nie, X. (2018). Detection and seg- mentation of mechanical damage in concrete with x-ray microtomography.Materials Characterization, 142:515–522

    2018

  38. [46]

    Lorenzoni, R., Curosu, I., Paciornik, S., Mechtcherine, V., Oppermann, M., and Silva, F. (2020). Semantic segmentation of the micro-structure of strain-hardening cement-based composites (shcc) by applying deep learning on micro-computed tomography scans.Cement and Concrete Com...

    2020

  39. [47]

    and Hutter, F

    Loshchilov, I. and Hutter, F. (2016). Sgdr: Stochastic gradient descent with warm restarts.arXiv preprint arXiv:1608.03983

    2016 arXiv

  40. [48]

    S., Zitouni, S., and Belâabes, S

    Meddah, M. S., Zitouni, S., and Belâabes, S. (2010). Effect of content and particle size distribution of coarse aggregate on the compressive strength of concrete.Construction and building materials, 24(4):505–512

    2010

  41. [49]

    Miao, Y., Wang, H., Zhang, Y., Feng, T., Yali, L., Liqun, L., Lin, J., Zeyu, L., and Jiang, J. (2024). A numerical model for predicting the drying shrinkage behavior of concrete under ultra-low relative humidity.Construction and Building Materials, 411:134459

    2024

  42. [50]

    Myong Lee, K., Buyukozturk, O., and Oumera, A. (1992). Fracture analysis of mortar-aggregate interfaces in concrete.Journal of engineering mechanics, 118(10):2031–2046

    1992

  43. [51]

    Naresh, K., Khan, K., Umer, R., and Cantwell, W. J. (2020). The use of x-ray computed tomography for design and process modeling of aerospace composites: A review.Materials & Design, 190:108553

    2020

  44. [52]

    and Protzel, P

    Neubert, P. and Protzel, P. (2014). Compact watershed and preemptive slic: On improving trade-offs of superpixel segmentation algorithms. In2014 22nd international conference on pattern recognition, pages 996–1001. IEEE

    2014

  45. [53]

    Niu, D., Wang, X., Han, X., Lian, L., Herzig, R., and Darrell, T. (2024). Unsupervised universal image segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 22744–22754. 30

    2024

  46. [54]

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019). Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32

    2019

  47. [55]

    Pecha, M., Cermak, M., Hapla, V., Horak, D., and Tomcala, J. (2017). Advanced approach of material region detections on fibre-reinforced concrete ct-scans.Advances in Electrical and Electronic Engineering, 15(2):223–229

    2017

  48. [56]

    Rao, G. A. and Raghu Prasad, B. (2004). Influence of type of aggregate and surface roughness on the interface fracture properties.Materials and Structures, 37:328–334

    2004

  49. [57]

    Ravi, N., Gabeur, V., Hu, Y.-T., Hu, R., Ryali, C., Ma, T., Khedr, H., Rädle, R., Rolland, C., Gustafson, L., et al. (2024). Sam 2: Segment anything in images and videos.arXiv preprint arXiv:2408.00714

    2024 arXiv

  50. [58]

    Ren, Q., Pacheco, J., and de Brito, J. (2023). Methods for the modelling of concrete mesostructures: a critical review.Construction and Building Materials, 408:133570

    2023

  51. [59]

    Ronneberger, O., Fischer, P., and Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. InMedical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part II...

    2015

  52. [60]

    E., Hinton, G

    Rumelhart, D. E., Hinton, G. E., and Williams, R. J. (1986). Learning representations by back- propagating errors.nature, 323(6088):533–536

    1986

  53. [61]

    K., Pradhan, S., and Barai, S

    Saha, S. K., Pradhan, S., and Barai, S. V. (2020). Use of machine learning based technique to x-ray microtomographic images of concrete for phase segmentation at meso-scale.Construction and Building Materials, 249:118744

    2020

  54. [62]

    L., Crumbie, A

    Scrivener, K. L., Crumbie, A. K., and Laugesen, P. (2004). The interfacial transition zone (itz) between cement paste and aggregate in concrete.Interface science, 12(4):411–421

    2004

  55. [63]

    J., Tan, C., and Wen, Y

    Shi, M., Hua, T., Yang, Z. J., Tan, C., and Wen, Y. (2024). Weakly supervised deep learning- based concrete aggregates automatic segmentation for assessing separation degree.Journal of Building Engineering, 82:108342

    2024

  56. [64]

    P., and Devabhaktuni, V

    Siddique, N., Paheding, S., Elkin, C. P., and Devabhaktuni, V. (2021). U-net and its variants for medical image segmentation: A review of theory and applications.Ieee Access, 9:82031–82057

    2021

  57. [65]

    C., Moreira, M

    Sleiman, H. C., Moreira, M. H., Tengattini, A., and Dal Pont, S. (2023). From tomographic imaging to numerical simulations: an open-source workflow for true morphology mesoscale fe meshes.RILEM Technical Letters, 8:158–164

    2023

  58. [66]

    Smith, S., Elsen, E., and De, S. (2020). On the generalization benefit of noise in stochastic gradient descent. InInternational Conference on Machine Learning, pages 9058–9067. PMLR

    2020

  59. [67]

    Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., and Salakhutdinov, R. (2014). Dropout: a simple way to prevent neural networks from overfitting.The journal of machine learning research, 15(1):1929–1958

    2014

  60. [68]

    Stamati, O., Roubin, E., Andò, E., and Malecot, Y. (2018). Phase segmentation of concrete x- ray tomographic images at meso-scale: Validation with neutron tomography.Cement and Concrete Composites, 88:8–16

    2018

  61. [69]

    Sutskever, I., Martens, J., Dahl, G., and Hinton, G. (2013). On the importance of initialization and momentum in deep learning. InInternational conference on machine learning, pages 1139–1147. PMLR. 31

    2013

  62. [70]

    Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., and Wojna, Z. (2016). Rethinking the inception architecture for computer vision. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 2818–2826

    2016

  63. [71]

    M., Enright, S., and Hurley, R

    Thakur, M. M., Enright, S., and Hurley, R. C. (2023). Phase segmentation in x-ray ct images of concrete with implications for mesoscale modeling.Construction and Building Materials, 403:133033

    2023

  64. [72]

    Tian, W., Cheng, X., Liu, Q., Yu, C., Gao, F., and Chi, Y. (2021). Meso-structure segmentation of concrete ct image based on mask and regional convolution neural network.Materials & Design, 208:109919

    2021

  65. [73]

    Tompson, J., Goroshin, R., Jain, A., LeCun, Y., and Bregler, C. (2015). Efficient object localization using convolutional networks. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 648–656

    2015

  66. [74]

    L., Nunez-Iglesias, J., Boulogne, F., Warner, J

    Van der Walt, S., Schönberger, J. L., Nunez-Iglesias, J., Boulogne, F., Warner, J. D., Yager, N., Gouillart, E., and Yu, T. (2014). scikit-image: image processing in python.PeerJ, 2:e453

    2014

  67. [75]

    Vásárhelyi, L., Kónya, Z., Kukovecz, Á., and Vajtai, R. (2020). Microcomputed tomography–based characterization of advanced materials: a review.Materials Today Advances, 8:100084

    2020

  68. [76]

    Wang, W., Su, C., and Zhang, H. (2022). Automatic segmentation of concrete aggregate using convolutional neural network.Automation in Construction, 134:104106

    2022

  69. [77]

    Wen, F., Li, W., Zhou, G., Li, H., Chen, T., and Zhang, K. (2024). A simplified numerical simulation of uniaxial compression for polyacrylonitrile fiber reinforced permeable concrete based on ct images. Construction and Building Materials, 411:134319

    2024

  70. [78]

    and Kulis, B

    Xia, X. and Kulis, B. (2017). W-net: A deep model for fully unsupervised image segmentation. arXiv preprint arXiv:1711.08506

    2017 arXiv

  71. [79]

    and Chun, P.-j

    Yamane, T. and Chun, P.-j. (2020). Crack detection from a concrete surface image based on semantic segmentation using deep learning.Journal of Advanced Concrete Technology, 18(9):493–504

    2020

  72. [80]

    Yang, Z., Ren, W., Sharma, R., McDonald, S., Mostafavi, M., Vertyagina, Y., and Marrow, T. (2017). In-situ x-ray computed tomography characterisation of 3d fracture evolution and image-based numerical homogenisation of concrete.Cement and Concrete Composites, 75:74–83

    2017

  73. [81]

    D., Krishnan, D., Taylor, G

    Zeiler, M. D., Krishnan, D., Taylor, G. W., and Fergus, R. (2010). Deconvolutional networks. In 2010 IEEE Computer Society Conference on computer vision and pattern recognition, pages 2528–2535. IEEE

    2010

  74. [82]

    H., Lura, P., Müller, H

    Zhang, P., Wittmann, F. H., Lura, P., Müller, H. S., Han, S., and Zhao, T. (2018). Application of neutron imaging to investigate fundamental aspects of durability of cement-based materials: A review. Cement and Concrete Research, 108:152–166

    2018

  75. [83]

    Zhou, S.-B., Shen, A.-Q., and Li, G.-F. (2015). Concrete image segmentation based on multiscale mathematic morphology operators and otsu method.Advances in Materials Science and Engineering, 2015(1):208473

    2015

  76. [84]

    Zhu, G., Wang, R., Liu, Y., Zhu, Z., Gao, C., Liu, L., and Sang, N. (2023). An adaptive post- processing network with the global-local aggregation for semantic segmentation.IEEE Transactions on Circuits and Systems for Video Technology. 32

    2023

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.