NormalView: tree species classification from backpack and aerial lidar data using geometric projections

Antero Kukko; Jesse Muhojoki; Josef Taher; Juha Hyypp\"a; Juho Korkeala; Klaara Salolahti; Matti Hyypp\"a

REVIEW 4 major objections 6 minor 1 cited by

NormalView: tree species classification from backpack and aerial lidar data using geometric projections

T0 review · 4 major / 6 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read NormalView shows that tree species can be classified from lidar point clouds using only geometry: images colored by surface normal direction reach 95.5% overall accuracy on dense mobile scans and 91.8% on airborne scans, rivaling models tha

desk verdict Solid, clearly reported empirical paper with a useful dataset release; the MLS superiority claim for normal-vector colouring needs repeated seeds before it is firm. read the letter →

arxiv 2512.05610 v2 pith:XD2QSSTL submitted 2025-12-05 cs.CV

Juho Korkeala , Jesse Muhojoki , Josef Taher , Klaara Salolahti , Matti Hyyppä , Antero Kukko , Juha Hyyppä This is my paper

classification cs.CV

keywords treespeciesclassificationmobilelaserscanningairbornepointcloudsgeometricprojectionssurfacenormalsdeeplearningmultispectrallidar

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 presents NormalView, a method that converts individual tree point clouds into 2D images colored by each point's surface normal direction, then feeds those images to a modern image-classification network. On dense mobile laser scanning (MLS) data covering seven boreal species, NormalView achieves 95.5% overall accuracy (94.8% macro-average), outperforming intensity-colored and silhouette-based projections. On airborne laser scanning (ALS) data with nine species, it reaches 91.8% overall accuracy, slightly behind the best multispectral-intensity model but ahead of single-channel intensity models. The authors argue that because the method relies only on geometry, it is sensor-agnostic and can work with any lidar scanner, and they release the MLS dataset (1,915 labeled segments) as a public benchmark.

What carries the argument

The key object is the normal-vector orthographic projection. For each point, a plane is fitted to its 20 nearest neighbors, the unit normal is oriented away from the trunk, and its x, y, z components become the red, green, and blue channels of projected images taken from five viewing angles (25 at inference) plus a 'sliced' view that reveals trunk and inner branches. This encodes local 3D orientation into a 2D image, letting a pretrained image-classification network exploit geometric structure that would otherwise be lost in projection. The paper compares this coloring with intensity-based and black-and-white projections to isolate the contribution of geometry.

What would settle it

Retrain the same models on multiple random 80/20 splits and compute confidence intervals; if the gap between NormalView and the intensity model on MLS data vanishes or flips across seeds, the claimed geometric advantage is not statistically reliable. Independently audit a random sample of matched tree pairs—if systematic label or matching errors are found, the absolute accuracy figures shrink.

Watch

Extended reading notes

Core claim

NormalView's central claim is that local geometric information, embedded as normal-vector RGB colors in multi-angle orthographic projections, is sufficient to distinguish tree species from high-density point clouds. On dense MLS data, this geometry-only approach is the strongest of the tested projection-based models, improving macro-average accuracy by four percentage points over black-and-white silhouettes. On ALS data, geometry alone is competitive with single-channel intensity, and only fusing three multispectral intensity channels clearly surpasses it—evidence that radiometric information matters most when point density and structural detail are lower. This demonstrates that projection-b

Load-bearing premise

The reported accuracies assume the ground-truth species labels and the 3-meter Hausdorff matching between scanned trees and reference stems are correct for both MLS and ALS; any mislabeled or mismatched tree directly biases the accuracy numbers, and a single 80/20 split with no repeated seeds leaves small model differences possibly within noise.

Editorial extensions

If this is right

Tree species can be classified from lidar using only geometry, so scanners without intensity or multispectral capability remain useful for species inventories.
Multispectral intensity from multiple scanners is beneficial on sparser airborne data, especially for minority species, where geometry-only performance drops more sharply.
Projection-based methods paired with modern image backbones are competitive with point-based 3D deep learning models on airborne data, contradicting the assumption that 3D architectures are necessary for high accuracy.
The released MLS dataset provides a public, labeled seven-species benchmark that future mobile-scanning classification work can build on.
The relative value of geometric versus radiometric features depends on point density: geometry dominates on dense ground-based scans, while radiometry helps when structure is incomplete.

Reading between the lines

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

NormalView's dependence on consistent trunk and branch geometry suggests its advantage will transfer best to ground-based scanning platforms; on low-density or heavily occluded point clouds, its benefit over simple silhouettes may shrink, and intensity or multispectral channels will matter more.
The same normal-coloring scheme could be applied to other point-cloud classification tasks, such as leaf/wood separation, dead-tree detection, or species classification from UAV and terrestrial scanners, wherever local surface orientation is estimable.
Combining normal vectors with bark-texture or eigenfeature channels, as the authors note in passing, could close the gap on airborne minority species like oak, which are currently the hardest to classify.
Because inference averages 50 images per tree, a practical extension would test whether fewer viewing angles or a single full-and-sliced pair can preserve most of the accuracy at lower computational cost for operational inventories.

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.

Desk Editor's Note

read the letter

Dear colleague,

Short version: this is a solid applied paper, worth sending to a serious referee. The new piece is modest—normal-vector colouring of orthographic projections, plugged into YOLOv11—but the experimental work is careful and the released MLS dataset is a genuine contribution.

The good stuff. The preprocessing pipeline is documented in unusual detail: 2 cm subsampling, SOR, CSF, manual cleaning, TreeIso, Hausdorff matching with a 3 m threshold. The field reference is the same as in Taher et al., and the benchmark comparison on ALS is to that paper, which is appropriate. Hyperparameters such as N=20, k=0.7 m, and image size were fixed before evaluation; the confusion matrices are internally consistent with the reported aggregate metrics. The finding that multispectral ALS intensity helps minority classes aligns with existing work, and the result that NormalView is competitive on ALS using only geometry is believable. Releasing the ~1900-tree MLS dataset and model weights is good practice and makes the paper reproducible.

The soft spot is statistical. Every configuration is run once on a single 80/20 split. On MLS, NormalView beats black-and-white by 1.5 percentage points in OA and 4.1 in MAA. With 333 test trees, and rowan n=9 and lime n=14, the MAA gap mostly comes from about four trees. YOLO training is stochastic—RandAugment, early stopping—so the ordering could easily flip across seeds. And because the ALS results do not favour NormalView over B&W, the MLS comparison is doing the work for the claim that normal-vector colouring adds value. The paper needs repeated seeds, confidence intervals, or a paired significance test before that claim is firm. The absence of label-error quantification for the field reference is a smaller caveat, not a deal-breaker.

This is a real weakness but not a fatal one. The absolute accuracies are high, the method is simple and sensor-agnostic, and the data release has standalone value. I would accept this for peer review, and I'd want a referee to push on the split variance. If that is addressed in revision, I'd be comfortable citing it in my own work.

Best, [you]

Referee Report

4 major / 6 minor

Summary. The paper proposes NormalView, a projection-based tree species classifier that colours multi-view orthographic projections by surface-normal RGB values and feeds them to a YOLOv11 image classifier. Experiments on high-density MLS data (7 species) and helicopter ALS data (7 and 9 species) compare NormalView with intensity-coloured and black-and-white projections. On MLS, NormalView reports OA/MAA of 95.5%/94.8%, outperforming intensity and silhouette models; on ALS, a three-channel intensity model achieves the best results (92.5%/84.9% for 9 species), with NormalView at 91.8%/79.1%. The authors release the MLS dataset and model weights.

Significance. The idea of encoding surface normals as RGB in multi-view orthographic projections is simple and sensible; if the comparative results hold, it offers a sensor-agnostic geometric baseline for projection-based tree species classification. The paper also provides a new public MLS tree-segment dataset and model weights, which is a concrete community resource. The ALS experiments systematically probe single- and multi-channel intensity, and the confusion matrices are presented in full. However, because the headline comparisons rest on a single split with small margins, the quantitative significance of the method's advantage is not yet established.

major comments (4)

[Sec. 3.1, Table 4] The central comparative claim—NormalView 'demonstrates the strongest performance' on MLS—is based on one deterministic 80/20 split. NormalView leads the black-and-white baseline by 1.5 pp OA and 4.1 pp MAA; with 333 test trees the OA gap is roughly 5 trees, and the MAA gap is mostly rowan (n=9) and lime (n=14) recall differences (Figure 5). Training uses RandAugment and early stopping (Sec. 2.4.2), so these margins are within plausible run-to-run noise. Please report repeated-seed results with means and intervals, or paired tests (e.g., McNemar/bootstrap) for the Table 4 comparisons.
[Sec. 3.1, Table 5] The ALS multispectral conclusion is similarly underdetermined. In Table 5, seven-species OAs lie within 1.0 pp and nine-species OAs within 1.3 pp across all ten variants. The MAA advantage of Channels 123 (90.6/84.9) over single-channel models is driven mainly by oak (n=13) and alder (n=27) in Figure 6, again with a single split and no variance estimates. State uncertainty or phrase the multispectral-benefit claim as suggestive rather than established.
[Sec. 2.4.1, Table 4] The NormalView vs black-and-white comparison is confounded by image post-processing. In Sec. 2.4.1, intensity and normal images receive a Gaussian 3x3 kernel (sigma=0.85), while WOP images receive no filter. Thus the Table 4 improvement of NV over B&W could be due to smoothing rather than normal-vector colouring. To support the statement in Sec. 3.3 that 'normal vector colouring improves classification performance', include an unsmoothed NV variant or apply the same filter to WOP.
[Sec. 3.3] The comparison to Taher et al. (2025) is based on 'a similar train-test split', not the exact partition used in the benchmark. If the test subsets differ, the stated equivalence (OA 92.5 vs 92.0, MAA 84.9 vs 85.1) is not a controlled comparison. Please use the same partition or explicitly qualify the comparison.

minor comments (6)

[Sec. 2.3, Table 2] The MLS Total row sums to 1904; the sum of the train entries is 1571, not the printed 1582, and the abstract states 1915 samples. The alder row (11 samples, 'not used in MLS models') is inconsistent with those totals. Please clarify and make all totals match.
[Abstract] Typo: 'assessing the decomposition of forest environments' likely should be 'composition'.
[Sec. 2.4.1 vs 2.4.2] Training images are created from 5 viewpoints, while inference uses 25 angles (50 images). Please state explicitly that multi-view inference uses more angles than training; as written it looks like a mismatch.
[Sec. 2.3] The 3 m Hausdorff matching and field reference are treated as ground truth. No estimate of label or matching error is given. All models share the reference, so this affects absolute accuracies more than comparisons, but it should be acknowledged as a limitation.
[Sec. 3.2, Figure 7] The image-size experiment is also single-run. Show variance or at least note that the observed peak at 512 is exploratory.
[Data availability] Typo: 'publically' should be 'publicly'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: NormalView's reported accuracies are empirical measurements on a held-out split, not outputs forced by fitted parameters or self-citations.

full rationale

The paper makes no formal derivation whose conclusion is identical to its premises. NormalView's inputs are point-cloud projections colored by estimated normal vectors; its outputs are tree-species labels, and the claimed 95.5%/94.8% MLS and 91.8%/79.1% ALS accuracies are measured on a held-out 20% test split. Hyperparameters such as N=20, k=0.7 m, image size, and the five-view projection setup are fixed before evaluation and are not fitted to test-set performance. The comparison with Taher et al. (2025) is a benchmark reference and a dataset-use statement, not a load-bearing argument that reduces NormalView's result to that citation; moreover the same-group benchmark is itself an empirical result with public data, so it is not an unverified uniqueness theorem. The paper's central comparative claim ('NormalView demonstrates the strongest performance on the MLS data') is an empirical observation that could in principle be contradicted by rerunning with different seeds; the lack of repeated splits and confidence intervals is a statistical robustness limitation, not circularity. No equation in Section 2.4 or elsewhere defines the method's success measure in terms of its own inputs, and no fitted parameter is renamed as a prediction. Accordingly, no circular step can be exhibited.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

Empirical deep-learning study with no theoretical derivation; the 'free parameters' are hyperparameters and preprocessing choices set by the authors, and the axioms are domain assumptions about data quality and the informativeness of the geometry-based representation.

free parameters (6)

N (normal estimation neighbourhood) = 20
Number of nearest points used for PCA plane fit when estimating surface normals (Sec 2.4.1); hand-set, affects the smoothness of normal-based colouring.
k (slice depth threshold) = 0.7 m
Threshold for excluding points in front of the trunk when creating 'sliced' projection images (Sec 2.4.1); chosen to reveal trunk/inner branches while tolerating trunk-position error.
Image size for model input = 512 (MLS) / 256 (ALS)
Training-time image resolution after bilinear scaling (Table 3); image-size experiment on MLS intensity images (Sec 3.2) selected 512 as best.
Gaussian smoothing kernel = 3x3, sigma=0.85
Applied to intensity and normal images to reduce discretisation noise (Sec 2.4.1).
Projection angle counts = 5 training / 25 inference
Number of rotations around the z-axis used to create projection images; the paper uses 5 views for training and aggregates 25 views at inference (Sec 2.4.1–2.4.2).
Training hyperparameters (learning rate, batch size, dropout, patience) = lr0=5e-4, batch 16/5, dropout 0.15, patience 30
AdamW settings from Table 3; not tuned to the test set but affect results.

assumptions (5)

domain assumption Field reference species labels are correct (ground truth for both MLS and ALS).
Section 2.3 describes reference data collected by FGI employees; any label errors directly distort reported accuracy.
domain assumption Hausdorff-distance nearest-neighbour matching with a 3 m threshold correctly associates MLS tree segments to reference trees.
Section 2.3; wrong matches would assign incorrect species labels to MLS segments.
domain assumption Normal vectors estimated via PCA over N=20 neighbours with outward orientation encode species-discriminative geometric structure.
Section 2.4.1; the entire NormalView approach assumes this representation preserves enough 3D structure after projection.
domain assumption Orthographic projections from 5 rotated viewpoints plus sliced images capture sufficient 3D tree structure for classification.
Section 2.4.1; the method deliberately loses one spatial dimension and relies on multi-view/slicing to compensate.
domain assumption ImageNet-pretrained YOLOv11 features transfer to tree projection images.
Section 2.4.2; the model is used as a pretrained classifier and fine-tuned; this is a standard but unproven-for-this-domain assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of NormalView: tree species classification from backpack and aerial lidar data using geometric projections." pith.science (2026). https://pith.science/paper/XD2QSSTL

@misc{pith2026251205610,
  author       = {Pith},
  title        = {Pith review of: NormalView: tree species classification from backpack and aerial lidar data using geometric projections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XD2QSSTL}},
  note         = {Machine review of arXiv:2512.05610}
}

read the original abstract

Laser scanning has proven to be an invaluable tool in assessing the decomposition of forest environments. Mobile laser scanning (MLS) has shown to be highly promising for extremely accurate, tree level inventory. In this study, we present NormalView, a projection-based deep learning method for classifying tree species from point cloud data. NormalView embeds local geometric information into two-dimensional projections, in the form of normal vector estimates, and uses the projections as inputs to an image classification network, YOLOv11. In addition, we inspected the effect of multispectral radiometric intensity information on classification performance. We trained and tested our model on high-density MLS data (7 species, ~5000 pts/m2), as well as high-density airborne laser scanning (ALS) data (9 species, >1000 pts/m2). On the MLS data, NormalView achieves an overall accuracy (macro-average accuracy) of 95.5 % (94.8 %), and 91.8 % (79.1 %) on the ALS data. We found that having intensity information from multiple scanners provides benefits in tree species classification, and the best model on the multispectral ALS dataset was a model using intensity information from all three channels of the multispectral ALS. This study demonstrates that projection-based methods, when enhanced with geometric information and coupled with state-of-the-art image classification backbones, can achieve exceptional results. Crucially, these methods rely only on geometric information, and thus are compatible with most sensors. Additionally, we publically release the MLS dataset used in the study, containing 1915 samples.

Figures

Figures reproduced from arXiv: 2512.05610 by the authors.

**Figure 2.** The same pine tree as captured by (a) the backpack mobile scanner, and (b) helicopter-ALS. The point counts of the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** (a)-(b) Two examples of images used in training the models. The images are WOP images created from the same birch [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (5 more)

**Figure 4.** Figure 4: Example images of the nine different tree species and image colouring methods used in model training. All samples are captured by HeliALS. For explanations for the colours, see section 2.4.1. In the WOP images, a pixel was simply coloured white, if there is a projected…

**Figure 5.** Figure 5: Species-wise metrics of the models trained on MLS data. (a) Species-wise precision, (b) species-wise recall, (c) [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]

**Figure 6.** Figure 6: Species-wise precision, recall, and F1-scores for models trained on ALS data with 7 (left) and 9 (right) species: (a)–(b) precision, (c)–(d) recall, (e)–(f) F1-scores [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

**Figure 7.** Figure 7: (a) Performance metrics of models trained on MLS data and varying image sizes. (b) Proportions of empty pixels in [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

**Figure 8.** Figure 8: Examples of MLS trees misclassified by at least one model. GT refers to ground truth, I, N and W refer to the [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Calibrated Tree-Neural Fusion for Fine-Grained Vegetation Community Classification
cs.LG 2026-07 conditional novelty 4.0 of 10

Calibrated EcoTreeFuseNet-Plus matches ExtraTrees on 29-class vegetation labels while cutting expected calibration error from 0.39 to 0.07 via temperature scaling and leakage-aware stacking.

Reference graph

Works this paper leans on

9 extracted references · 3 canonical work pages · cited by 1 Pith paper

[6]

Kivinen, S., Koivisto, E., Keski-Saari, S., Poikolainen, L., Tanhuanpää, T., Kuzmin, A., Viinikka, A., Heikkinen, R

doi:10.3390/rs4040950. Kivinen, S., Koivisto, E., Keski-Saari, S., Poikolainen, L., Tanhuanpää, T., Kuzmin, A., Viinikka, A., Heikkinen, R. K., Pykälä, J., Virkkala, R., Vihervaara, P., and Kumpula, T. A keystone species, european aspen (populus tremula l.), in boreal forests: Eco- logical role, knowledge needs and mapping using remote sensing.Forest Ecol...

arXiv 2020
[7]

doi:10.1016/j.jag.2025.104502

ISSN 1569-8432. doi:10.1016/j.jag.2025.104502. Tran, D., Bourdev, L., Fergus, R., Torresani, L., and Paluri, M. Learning spatiotemporal features with 3d convolutional networks. InProceedings of the IEEE international conference on computer vision, pages 4489–4497, 2015. Wang, Y ., Sun, Y ., Liu, Z., Sarma, S. E., Bronstein, M. M., and Solomon, J. M. Dy- n...

arXiv 2025
[1112]

Yu, X., Hyyppä, J., Holopainen, M., and Vastaranta, M

doi:10.14358/PERS.72.12.1339. Yu, X., Hyyppä, J., Holopainen, M., and Vastaranta, M. Comparison of area-based and in- dividual tree-based methods for predicting plot-level forest attributes.Remote Sensing, 2(6):1481–1495, 2010. doi:10.3390/rs2061481. Yu, X., Hyyppä, J., Litkey, P., Kaartinen, H., Vastaranta, M., and Holopainen, M. Single- sensor solution ...

work page doi:10.14358/pers.72.12.1339 2010
[1995]

Cubuk, E

doi:10.1007/BF00994018. Cubuk, E. D., Zoph, B., Shlens, J., and Le, Q. V . Randaugment: Practical automated data augmentation with a reduced search space. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 3008– 3017, June 2020. doi:10.48550/arXiv.1909.13719. Dalponte, M., Bruzzone, L., and Gianel...

arXiv doi:10.48550/arxiv.1909.13719 2020
[2012]

CloudCompare development team

doi:10.5849/forsci.12-003. CloudCompare development team. Cloudcompare, version 2.13.1 [software], 2024. URL http://www.cloudcompare.org/. Cohen, J. A coefficient of agreement for nominal scales.Educational and Psychological Measurement, 20(1):37–46, 1960. doi:10.1177/001316446002000104. Cortes, C. and Vapnik, V . Support-vector networks.Machine learning,...

work page doi:10.5849/forsci.12-003 2024
[2014]

Izenman, A

doi:10.48550/arXiv.1404.1869. Izenman, A. J. Linear discriminant analysis. InModern multivariate statistical techniques: regression, classification, and manifold learning, pages 237–280. Springer, 2013. Jocher, G., Qiu, J., and Chaurasia, A. Ultralytics YOLO, Version 8.0.0, Jan. 2023. URL https://github.com/ultralytics/ultralytics. Kaartinen, H., Hyyppä, ...

arXiv doi:10.48550/arxiv.1404.1869 2013
[2016]

Zhao, H., Jiang, L., Jia, J., Torr, P., and Koltun, V

doi:10.3390/rs8060501. Zhao, H., Jiang, L., Jia, J., Torr, P., and Koltun, V . Point transformer. InPro- ceedings - 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021, Proceedings of the IEEE International Conference on Computer Vision, pages 16239–16248. Institute of Electrical and Electronics Engineers Inc., 2021. doi:10.1109/ICCV48922...

arXiv 2021
[2020]

Hyyppä, M., Salolahti, K., Hyyppä, E., Yu, X., Taher, J., Matikainen, L., Litkey, P., Lehtomäki, M., Hakala, T., Kaartinen, H., Hyyppä, J., and Kukko, A

doi:10.3390/rs12203327. Hyyppä, M., Salolahti, K., Hyyppä, E., Yu, X., Taher, J., Matikainen, L., Litkey, P., Lehtomäki, M., Hakala, T., Kaartinen, H., Hyyppä, J., and Kukko, A. Multispectral airborne laser scanning for species classification in espoonlahti (MS-ALS-SPECIES) [dataset]. Zenodo, 2025. doi: 10.5281/zenodo.17077256. Iandola, F. N., Moskewicz, ...

arXiv 2025

Show all 9 references

[2025]

Bradski, G

doi:10.1002/rse2.70014. Bradski, G. The OpenCV Library.Dr. Dobb’s Journal of Software Tools, 2000. Breiman, L. Random forests.Machine learning, 45(1):5–32, 2001. doi:10.1023/A:1010933404324. Chirici, G., McRoberts, R. E., Winter, S., Bertini, R., Brändli, U.-B., Asensio, I. A....

2000 doi

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [6]

Kivinen, S., Koivisto, E., Keski-Saari, S., Poikolainen, L., Tanhuanpää, T., Kuzmin, A., Viinikka, A., Heikkinen, R

doi:10.3390/rs4040950. Kivinen, S., Koivisto, E., Keski-Saari, S., Poikolainen, L., Tanhuanpää, T., Kuzmin, A., Viinikka, A., Heikkinen, R. K., Pykälä, J., Virkkala, R., Vihervaara, P., and Kumpula, T. A keystone species, european aspen (populus tremula l.), in boreal forests: Eco- logical role, knowledge needs and mapping using remote sensing.Forest Ecol...

arXiv 2020

[2] [7]

doi:10.1016/j.jag.2025.104502

ISSN 1569-8432. doi:10.1016/j.jag.2025.104502. Tran, D., Bourdev, L., Fergus, R., Torresani, L., and Paluri, M. Learning spatiotemporal features with 3d convolutional networks. InProceedings of the IEEE international conference on computer vision, pages 4489–4497, 2015. Wang, Y ., Sun, Y ., Liu, Z., Sarma, S. E., Bronstein, M. M., and Solomon, J. M. Dy- n...

arXiv 2025

[3] [1112]

Yu, X., Hyyppä, J., Holopainen, M., and Vastaranta, M

doi:10.14358/PERS.72.12.1339. Yu, X., Hyyppä, J., Holopainen, M., and Vastaranta, M. Comparison of area-based and in- dividual tree-based methods for predicting plot-level forest attributes.Remote Sensing, 2(6):1481–1495, 2010. doi:10.3390/rs2061481. Yu, X., Hyyppä, J., Litkey, P., Kaartinen, H., Vastaranta, M., and Holopainen, M. Single- sensor solution ...

work page doi:10.14358/pers.72.12.1339 2010

[4] [1995]

Cubuk, E

doi:10.1007/BF00994018. Cubuk, E. D., Zoph, B., Shlens, J., and Le, Q. V . Randaugment: Practical automated data augmentation with a reduced search space. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 3008– 3017, June 2020. doi:10.48550/arXiv.1909.13719. Dalponte, M., Bruzzone, L., and Gianel...

arXiv doi:10.48550/arxiv.1909.13719 2020

[5] [2012]

CloudCompare development team

doi:10.5849/forsci.12-003. CloudCompare development team. Cloudcompare, version 2.13.1 [software], 2024. URL http://www.cloudcompare.org/. Cohen, J. A coefficient of agreement for nominal scales.Educational and Psychological Measurement, 20(1):37–46, 1960. doi:10.1177/001316446002000104. Cortes, C. and Vapnik, V . Support-vector networks.Machine learning,...

work page doi:10.5849/forsci.12-003 2024

[6] [2014]

Izenman, A

doi:10.48550/arXiv.1404.1869. Izenman, A. J. Linear discriminant analysis. InModern multivariate statistical techniques: regression, classification, and manifold learning, pages 237–280. Springer, 2013. Jocher, G., Qiu, J., and Chaurasia, A. Ultralytics YOLO, Version 8.0.0, Jan. 2023. URL https://github.com/ultralytics/ultralytics. Kaartinen, H., Hyyppä, ...

arXiv doi:10.48550/arxiv.1404.1869 2013

[7] [2016]

Zhao, H., Jiang, L., Jia, J., Torr, P., and Koltun, V

doi:10.3390/rs8060501. Zhao, H., Jiang, L., Jia, J., Torr, P., and Koltun, V . Point transformer. InPro- ceedings - 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021, Proceedings of the IEEE International Conference on Computer Vision, pages 16239–16248. Institute of Electrical and Electronics Engineers Inc., 2021. doi:10.1109/ICCV48922...

arXiv 2021

[8] [2020]

Hyyppä, M., Salolahti, K., Hyyppä, E., Yu, X., Taher, J., Matikainen, L., Litkey, P., Lehtomäki, M., Hakala, T., Kaartinen, H., Hyyppä, J., and Kukko, A

doi:10.3390/rs12203327. Hyyppä, M., Salolahti, K., Hyyppä, E., Yu, X., Taher, J., Matikainen, L., Litkey, P., Lehtomäki, M., Hakala, T., Kaartinen, H., Hyyppä, J., and Kukko, A. Multispectral airborne laser scanning for species classification in espoonlahti (MS-ALS-SPECIES) [dataset]. Zenodo, 2025. doi: 10.5281/zenodo.17077256. Iandola, F. N., Moskewicz, ...

arXiv 2025