Pith. sign in

REVIEW 3 major objections 6 minor 48 references

SilvaScenes: Tree Detection and Species Classification from Under-Canopy Images in Natural Forests

T0 review · 3 major / 6 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read The paper introduces SilvaScenes, a benchmark dataset for under-canopy tree species instance segmentation, and shows that while locating tree trunks is feasible, identifying their species remains a largely unsolved problem.

desk verdict Valuable under-canopy tree species segmentation dataset with a serious abstract/full-text inconsistency and unmeasured label noise; needs revision but deserves refereeing. read the letter →

arxiv 2510.09458 v2 pith:S5YMW3ZZ submitted 2025-10-10 cs.CV cs.AIcs.LGcs.RO

David-Alexandre Duclos , William Guimont-Martin , Gabriel Jeanson , Arthur Larochelle-Tremblay , Martine Lapointe , Théo Defosse , Frédéric Moore , Philippe Nolet
show 2 more authors
François Pomerleau Philippe Giguère
This is my paper · ORCID
classification cs.CVcs.AIcs.LGcs.RO
keywords under-canopyimagerytreespeciesclassificationinstancesegmentationforestryroboticsbenchmarkdatasetimbalanceimageresolutionMask2Former
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 introduces SilvaScenes, a dataset of 172 under-canopy images from natural forests, with 1,476 trees across 24 species annotated by forestry experts. Its central claim is that while locating and segmenting tree trunks is feasible with modern deep learning (best mAP 67.65%), identifying the species of each segmented trunk remains a hard, unresolved problem (best mAP 35.69%). The authors establish this by benchmarking several instance segmentation models under a common protocol, and they show that image resolution, species imbalance, and occlusion are key factors. A reader should care because forestry automation and inventory depend on exactly this perception capability.

What carries the argument

The load-bearing element is the dataset itself: 172 high-resolution under-canopy images collected across five bioclimatic domains in Quebec, with trunk instance masks and fine-grained species labels assigned in situ by forestry experts. Annotation rules — trunk-only masks, a 16-pixel median-width threshold for labeling, forking trunks counted as separate trees, and grouping of rare or unidentifiable trees into an 'Other' class — define the task. The benchmark uses Mask2Former with Swin backbones and YOLO variants, trained with focal loss under five-fold stratified cross-validation, and the resolution ablation (downsampling from 1.6 MP to 0.1 MP) is the key diagnostic showing a power-law impr

What would settle it

Re-annotate a random subset of SilvaScenes images with multiple independent forestry experts and measure label agreement; if agreement is low, then a large part of the species-classification gap is annotation noise rather than perceptual difficulty. Alternatively, train the same models on a carefully balanced subset and see if mAP rises sharply, which would confirm species imbalance as the dominant factor.

Watch

Extended reading notes

Core claim

The central discovery is that under-canopy species-aware instance segmentation is a distinct and currently unsolved task, and SilvaScenes is offered as the benchmark to measure progress on it. The paper shows a large gap between class-agnostic trunk segmentation and species classification, with the best model reaching 67.65% mAP for the former and only 35.69% for the latter. It also finds that doubling image resolution improves mAP by about 6%, that confusion is concentrated among visually similar species such as spruces, and that the common practice of downsampling images may be holding back performance.

Load-bearing premise

The ground-truth species labels are assigned by forestry experts but with no measured inter-expert agreement, so if expert labels are noisy or biased, the benchmark's reported difficulty could be overstated.

Editorial extensions

If this is right

  • Species-aware under-canopy segmentation is an open problem; current models leave large room for improvement (best mAP 35.69%).
  • Higher-resolution imagery is a reliable lever: mAP increases roughly 6% per doubling of resolution, suggesting the field should move beyond standard downsampled inputs.
  • Species imbalance and occlusion are the dominant error sources; the confusion matrix shows frequent confusion between red/sugar maples and among spruces.
  • The dataset provides a public benchmark for future work on forestry perception, semantic SLAM, and precision forestry.
  • Binary trunk segmentation results show that detection itself is close to solved in these conditions, so the bottleneck is fine-grained species discrimination.

Reading between the lines

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

  • Because ground-truth species labels come from a single expert per site with no reported inter-expert agreement, part of the 35.69% mAP ceiling could reflect label ambiguity rather than intrinsic visual difficulty; a multi-annotator reliability study would disambiguate this.
  • The resolution scaling trend suggests that the current practice of downsampling to ~1 MP is a major bottleneck; if the trend holds, using full 100 MP images or learned super-resolution could close a substantial fraction of the gap.
  • The confusion pattern between deciduous and coniferous species (only 8% of errors) hints that a coarse 'tree type' classifier is nearly solved, and fine-grained species ID might benefit from additional cues like bark texture, context, or temporal information.
  • The dataset's explicit avoidance of duplicate trees across images controls data leakage, but it also means models never see the same tree from multiple viewpoints — a condition likely in real robotic deployments, so a multi-view extension would be a natural next benchmark.
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

3 major / 6 minor

Summary. The manuscript introduces SilvaScenes, an under-canopy RGB dataset of 172 images collected across five bioclimatic domains in Quebec, containing 1,476 individually annotated trees from 24 species (plus an Unknown class), with instance segmentation masks and forestry-expert species labels. The authors benchmark six model variants (Mask2Former with Swin-Small/Swin-Large, YOLOv11 Small/X-Large, YOLOv12 Small/X-Large) under a stratified five-fold cross-validation protocol using standard instance-segmentation metrics. The best species-aware mAP is 35.69% for Mask2Former with Swin-Large, while the full-text abstract reports a binary trunk-segmentation mAP of 67.65%. The authors conclude that trunk segmentation is feasible but fine-grained species classification remains a substantial challenge, and that higher image resolution yields consistent performance gains.

Significance. If the dataset and labels are trustworthy, SilvaScenes fills a real gap: existing ground-level datasets either lack species labels, cover only a handful of visually distinct species, or were collected in urban or snowy settings with lower clutter. The benchmark is reasonably designed — standard metrics, multi-architecture comparison, stratified cross-validation, and a useful confusion-matrix analysis. The resolution-scaling experiment is a valuable addition. However, the current manuscript contains a serious inconsistency in headline numbers between the arXiv abstract (1,421 trees, 28 species, mAP 69.9/39.2) and the full text (1,476 trees, 24 species, mAP 67.65/35.69), and the species labels — the core scientific payload — are not backed by any inter-expert agreement or label-noise analysis. These issues need to be resolved before the quantitative claims can be accepted.

major comments (3)
  1. [Abstract vs. full text; Table II] The headline numbers are inconsistent across versions of the same manuscript. The arXiv abstract reports 1,421 trees, 28 species, mAP 69.9%/39.2%, whereas the full-text abstract reports 1,476 trees, 24 species, mAP 67.65%/35.69%. Table II supports 35.69% as the best species-aware mAP for M2F-Large, but no table or protocol is provided for the binary-trunk mAP of 67.65% (or 69.9%). The authors must use a single consistent set of numbers and explicitly state how the binary segmentation mAP was computed, since it is not derivable from the reported species-aware experiments.
  2. [Section III-D, item 1; Section V] Ground-truth species labels are load-bearing for the central claim that 'species classification remains a significant challenge.' The text states that 'ground truth for most of the data was obtained in situ by a forestry expert,' but does not specify how the remaining data were labeled, nor does it provide inter-expert agreement, label-noise estimates, or any validation of the expert identifications. Confusion patterns such as BBP/PAB in Fig. 3 could be caused by bark detachment in the field, but they could also reflect label error. The authors should report the number of trees/images labeled by the primary expert versus any other process, provide an independent expert audit on a subset, and discuss the expected impact of label noise on the 35.69% mAP figure.
  3. [Section IV-B; Table I] The benchmark trains on 21 classes because four species with fewer than ten trees are merged with Unknown into an 'Other' class. This is methodologically defensible, but the paper presents SilvaScenes as a 24-species dataset. The result is that per-species performance for those four species cannot be measured, and 'Other' conflates rare species with damaged/unidentifiable trees, making the task harder. The authors should identify the four merged species explicitly and report per-species AP for all 24 species (even if low-confidence), or clearly state that the benchmark covers 24 species with only 21 trainable classes.
minor comments (6)
  1. [Introduction] Typo: 'UA V' should be 'UAV' in Section I.
  2. [Section IV-A] Please specify the exact YOLO variants used (e.g., YOLOv11n/s/m/l/x and YOLOv12n/s/m/l/x) and the source versions for reproducibility.
  3. [Section V / Fig. 5] The 'Multi → Binary' and 'Binary → Binary' curves rely on binary labels that are not defined in detail. State whether binary masks are simply the union of all species masks and whether the same five-fold splits are used. Also define 'IQR over five folds' more precisely in the caption.
  4. [Table I] The species code 'Pensylvanica' should be 'pensylvanica' for the epithet, and the Unknown class is missing from the taxonomic grouping. Consider adding a total column to make the per-domain sums easier to verify.
  5. [Dataset availability] The manuscript says the dataset and code 'will be made available' but gives no currently accessible link, license, or data sheet. A reviewer cannot verify the central claims without access; please provide at least a clear availability statement with a persistent identifier.
  6. [References] Reference [47] is cited to support the claim about typical species imbalance in natural forests, but it concerns citizen-science UAV data; a more direct forestry reference would strengthen that sentence.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SilvaScenes is a dataset-and-benchmark paper whose headline numbers are measured on held-out folds, not derived from fitted parameters or self-referential definitions.

full rationale

The paper's central claims are empirical: it introduces a dataset and reports instance-segmentation and species-classification metrics from five-fold cross-validation of standard models (Mask2Former, YOLOv11/12). The mAP values, confusion matrices, and resolution-ablation curves are measurements on held-out data, not quantities that reduce by construction to the dataset curation rules. The annotation guidelines (median width >= 16 px, Unknown class, grouping rare species into Other) are explicit preprocessing choices; they shape the benchmark task but do not define the resulting model performance. No fitted parameter is renamed as a prediction, and no claimed result is derived from the definitions of the targets. Self-citations appear (e.g., BarkNet [30], CanaTree100 [28]), but they are used as external related work/datasets and are not load-bearing justifications for the paper's own benchmark outcomes. The concern about expert-label reliability is a legitimate data-quality limitation, not a circularity: even if labels were noisy, the reported numbers would still be empirical measurements of performance against those labels, not tautological derivations. Thus the appropriate circularity score is 0.

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

The paper makes no free-parameter fits or invented entities. The main implicit assumptions are annotation-quality and curation decisions. These are transparently stated but not externally validated; they shape the benchmark and should be disclosed when comparing to other datasets.

assumptions (3)
  • domain assumption Ground truth species labels are assumed correct because they come from in situ forestry experts.
    Section III-D states ground truth was obtained in situ by a forestry expert, but no inter-expert agreement is reported. The reliability of the benchmark hinges on this assumption.
  • ad hoc to paper Trees are labeled only if their median trunk width is at least 16 px; obstructed segments are labeled only if shape can be inferred.
    This annotation rule (Section III-D.5) defines what is annotated and may bias the task toward more visible trees. It is a paper-specific choice, not a standard or externally validated threshold.
  • ad hoc to paper Four rare species are grouped with Unknown into an 'Other' class during experiments.
    Section IV-B explains that species with fewer than ten trees are merged into Other. This changes the classification task and can make the dataset appear easier or harder depending on how the class is treated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SilvaScenes: Tree Detection and Species Classification from Under-Canopy Images in Natural Forests." pith.science (2026). https://pith.science/paper/S5YMW3ZZ

@misc{pith2026251009458,
  author       = {Pith},
  title        = {Pith review of: SilvaScenes: Tree Detection and Species Classification from Under-Canopy Images in Natural Forests},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S5YMW3ZZ}},
  note         = {Machine review of arXiv:2510.09458}
}
read the original abstract

Interest in forestry automation is growing alongside rapid advances in deep learning. In particular, tree detection and taxonomic classification are seen as core tasks required for automating field surveys and forestry equipment. These operations must often be performed in under-canopy settings, which pose challenging conditions for perception systems, including heavy occlusion, variable lighting, and dense vegetation. Despite this necessity, current work has yet to properly establish the feasibility of simultaneously executing tree detection and taxonomic classification in natural forests, as available datasets primarily focus on urban settings or on a limited number of species. To address this gap, we present SilvaScenes, a benchmark dataset for instance segmentation of tree species from under-canopy images in natural forests. Collected across five bioclimatic domains in Quebec, Canada, our dataset features 1421 trees from 28 species, with segmentation masks for pixel-precise tree trunk detection and fine-grained species annotations from forestry experts. We demonstrate the relevance and difficult nature of SilvaScenes by evaluating modern deep learning approaches, showing that while trunk segmentation is feasible, with a top mean average precision (mAP) of 69.9% and mean average recall (mAR) of 76.4%, species-aware segmentation remains a significant challenge with an mAP and an mAR of only 39.2% and 68.6%, respectively. Alongside additional experiments, we highlight key challenges, namely that species imbalance and tree occlusion figure among the most pressing issues for precise segmentation and identification. Meanwhile, higher image resolutions contribute to significant performance gains and will likely prove fundamental to these tasks moving forward. Our dataset, source code, and models will be made available at https://github.com/norlab-ulaval/SilvaScenes.

Figures

Figures reproduced from arXiv: 2510.09458 by the authors.

Figure 1
Figure 1. Example of an annotated image in our dataset, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Statistics of SilvaScenes. (a) Number of trees per image. (b) Number of species per image. (c) Log-scale distribution of tree width in our images [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Examples of instance segmentation predictions with [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figures from the paper (1 more)
Figure 3
Figure 3. Figure 3: Confusion matrix of M2F-Large over five folds. Results are row-normalized and expressed in percentages. Species are split into deciduous, coniferous, and Other, and grouped to highlight inter- and intra-genus confusion. Qualitative results for M2F-Large are presented i…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 2 linked inside Pith

  1. [1]

    Digiforests: a Longitudinal Lidar Dataset for Forestry Robotics,

    M. V . R. Malladi, N. Chebrolu, I. Scacchetti, et al., “Digiforests: a Longitudinal Lidar Dataset for Forestry Robotics,” in IEEE International Conference on Robotics and Automation (ICRA) , 2025, pp. 1459– 1466

    2025

  2. [2]

    A review of UAS-based estimation of forest traits and characteristics in landscape ecology,

    A. I. Spiers, V . M. Scholl, J. McGlinchy, J. Balch, and M. E. Cattau, “A review of UAS-based estimation of forest traits and characteristics in landscape ecology,” Landscape Ecology, vol. 40, no. 2, p. 29, 2025

    2025

  3. [3]

    Build- ing Forest Inventories With Autonomous Legged Robots—System, Lessons, and Challenges Ahead,

    M. Mattamala, N. Chebrolu, J. Frey, et al. , “Build- ing Forest Inventories With Autonomous Legged Robots—System, Lessons, and Challenges Ahead,” IEEE Transactions on Field Robotics (T-FR) , vol. 2, pp. 418–436, 2025

    2025

  4. [4]

    Tree Instance Segmentation and Traits Estimation for Forestry Environments Exploiting LiDAR Data Col- lected by Mobile Robots,

    M. V . R. Malladi, T. Guadagnino, L. Lobefaro, et al., “Tree Instance Segmentation and Traits Estimation for Forestry Environments Exploiting LiDAR Data Col- lected by Mobile Robots,” in IEEE International Con- ference on Robotics and Automation (ICRA) , 2024, pp. 17 933–17 940

    2024

  5. [5]

    Robotic Precision Harvesting: Mapping, Localization, Plan- ning and Control for a Legged Tree Harvester,

    E. Jelavic, D. Jud, P. Egli, and M. Hutter, “Robotic Precision Harvesting: Mapping, Localization, Plan- ning and Control for a Legged Tree Harvester,” Field Robotics, vol. 2, pp. 1386–1431, 2022

    2022

  6. [6]

    SLOAM: Semantic Lidar Odometry and Mapping for Forest Inventory,

    S. W. Chen, G. V . Nardari, E. S. Lee, et al., “SLOAM: Semantic Lidar Odometry and Mapping for Forest Inventory,” IEEE Robotics and Automation Letters (RA-L), vol. 5, no. 2, pp. 612–619, 2020

    2020

  7. [7]

    A Stereo Visual- Inertial SLAM Algorithm with Point-Line Fusion and Semantic Optimization for Forest Environments,

    B. Liu, H. Liu, Y . Xing, et al. , “A Stereo Visual- Inertial SLAM Algorithm with Point-Line Fusion and Semantic Optimization for Forest Environments,” Forests, vol. 16, no. 2, p. 335, 2025

    2025

  8. [8]

    OpenForest: A data catalogue for ma- chine learning in forest monitoring,

    A. Ouaknine, T. Kattenborn, E. Lalibert ´e, and D. Rolnick, “OpenForest: A data catalogue for ma- chine learning in forest monitoring,” arXiv preprint arXiv:2311.00277, 2023

    arXiv 2023

Show all 48 references
  1. [9]

    FinnWoodlands Dataset,

    J. Lagos, U. Lempi ¨o, and E. Rahtu, “FinnWoodlands Dataset,” in Scandinavian Conference on Image Anal- ysis (SCIA), Springer, 2023, pp. 95–110

    2023

  2. [10]

    Classification of tree species and stock volume estimation in ground forest images using Deep Learning,

    J. Liu, X. Wang, and T. Wang, “Classification of tree species and stock volume estimation in ground forest images using Deep Learning,” Computers and Electronics in Agriculture, vol. 166, p. 105 012, 2019

    2019

  3. [11]

    Automated identification of tree species from images of the bark, leaves and needles,

    S. Fiel and R. Sablatnig, “Automated identification of tree species from images of the bark, leaves and needles,” in 16th Computer Vision Winter Workshop , 2011

    2011

  4. [12]

    A Review: Tree Species Classification Based on Remote Sensing Data and Classic Deep Learning- Based Methods,

    L. Zhong, Z. Dai, P. Fang, Y . Cao, and L. Wang, “A Review: Tree Species Classification Based on Remote Sensing Data and Classic Deep Learning- Based Methods,” Forests, vol. 15, no. 5, p. 852, 2024

    2024

  5. [13]

    Compar- ison of Backpack, Handheld, Under-Canopy UA V, and Above-Canopy UA V Laser Scanning for Field Reference Data Collection in Boreal Forests,

    E. Hyypp ¨a, X. Yu, H. Kaartinen, et al. , “Compar- ison of Backpack, Handheld, Under-Canopy UA V, and Above-Canopy UA V Laser Scanning for Field Reference Data Collection in Boreal Forests,” Remote Sensing, vol. 12, no. 20, p. 3327, 2020

    2020

  6. [14]

    Mapping Forest Parameters to Model the Mobility of Terrain Vehicles,

    T. Mikita, M. Rybansky, D. Krauskov ´a, F. Dohnal, O. Vystav ˇel, and S. Hollmannov ´a, “Mapping Forest Parameters to Model the Mobility of Terrain Vehicles,” Forests, vol. 15, no. 11, p. 1882, 2024

    2024

  7. [15]

    TreeScope: An Agricultural Robotics Dataset for LiDAR-Based Mapping of Trees in Forests and Orchards,

    D. Cheng, F. Cladera, A. Prabhu, et al., “TreeScope: An Agricultural Robotics Dataset for LiDAR-Based Mapping of Trees in Forests and Orchards,” in IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 14 860–14 866

    2024

  8. [16]

    SegmentAnyTree: A sensor and platform agnostic deep learning model for tree segmentation using laser scanning data,

    M. Wielgosz, S. Puliti, B. Xiang, K. Schindler, and R. Astrup, “SegmentAnyTree: A sensor and platform agnostic deep learning model for tree segmentation using laser scanning data,” Remote Sensing of Envi- ronment, vol. 313, p. 114 367, 2024

    2024

  9. [17]

    Bench- marking tree species classification from proximally sensed laser scanning data: Introducing the FOR- species20K dataset,

    S. Puliti, E. R. Lines, J. M ¨ullerov´a, et al. , “Bench- marking tree species classification from proximally sensed laser scanning data: Introducing the FOR- species20K dataset,” Methods in Ecology and Evolu- tion, vol. 16, no. 4, pp. 801–818, 2025

    2025

  10. [18]

    Automated pruning decisions in dormant sweet cherry canopies using instance segmentation,

    D. Borrenpohl and M. Karkee, “Automated pruning decisions in dormant sweet cherry canopies using instance segmentation,” Computers and Electronics in Agriculture, vol. 207, p. 107 716, 2023

    2023

  11. [19]

    Towards smart pruning: ViNet, a deep- learning approach for grapevine structure estimation,

    T. Gentilhomme, M. Villamizar, J. Corre, and J.-M. Odobez, “Towards smart pruning: ViNet, a deep- learning approach for grapevine structure estimation,” Computers and Electronics in Agriculture , vol. 207, p. 107 736, 2023

    2023

  12. [20]

    Learning Occluded Branch Depth Maps in Forest Environments Using RGB-D Images,

    C. Geckeler, E. Aucone, Y . Schnider, et al., “Learning Occluded Branch Depth Maps in Forest Environments Using RGB-D Images,” IEEE Robotics and Automa- tion Letters (RA-L) , vol. 9, no. 3, pp. 2439–2446, 2024

    2024

  13. [21]

    Robotic har- vesting of the occluded fruits with a precise shape and position reconstruction approach,

    L. Gong, W. Wang, T. Wang, and C. Liu, “Robotic har- vesting of the occluded fruits with a precise shape and position reconstruction approach,” Journal of Field Robotics, vol. 39, no. 1, pp. 69–84, 2022

    2022

  14. [22]

    Instance Segmentation for Autonomous Log Grasping in Forestry Operations,

    J.-M. Fortin, O. Gamache, V . Grondin, F. Pomer- leau, and P. Gigu `ere, “Instance Segmentation for Autonomous Log Grasping in Forestry Operations,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , 2022, pp. 6064–6071

    2022

  15. [23]

    TimberVision: A Multi-Task Dataset and Framework for Log-Component Segmentation and Tracking in Autonomous Forestry Operations,

    D. Steininger, J. Simon, A. Trondl, and M. Murschitz, “TimberVision: A Multi-Task Dataset and Framework for Log-Component Segmentation and Tracking in Autonomous Forestry Operations,” in IEEE/CVF Win- ter Conference on Applications of Computer Vision (WACV), 2025, pp. 5601–5610

    2025

  16. [24]

    A real time LiDAR- Visual-Inertial object level semantic SLAM for forest environments,

    H. Liu, G. Xu, B. Liu, et al. , “A real time LiDAR- Visual-Inertial object level semantic SLAM for forest environments,” ISPRS Journal of Photogrammetry and Remote Sensing (P&RS) , vol. 219, pp. 71–90, 2025

    2025

  17. [25]

    WildScenes: A benchmark for 2D and 3D semantic segmentation in large-scale natural environments,

    K. Vidanapathirana, J. Knights, S. Hausler, et al. , “WildScenes: A benchmark for 2D and 3D semantic segmentation in large-scale natural environments,”The International Journal of Robotics Research (IJRR) , vol. 44, no. 4, pp. 532–549, 2025

    2025

  18. [26]

    The GOOSE Dataset for Perception in Unstructured Environments,

    P. Mortimer, R. Hagmanns, M. Granero, T. Luet- tel, J. Petereit, and H.-J. Wuensche, “The GOOSE Dataset for Perception in Unstructured Environments,” in IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 14 838–14 844

    2024

  19. [27]

    Visible and Thermal Image-Based Trunk Detection with Deep Learning for Forestry Mobile Robotics,

    D. Q. Da Silva, F. N. Dos Santos, A. J. Sousa, and V . Filipe, “Visible and Thermal Image-Based Trunk Detection with Deep Learning for Forestry Mobile Robotics,” Journal of Imaging , vol. 7, no. 9, p. 176, 2021

    2021

  20. [28]

    Tree detection and diameter estimation based on deep learning,

    V . Grondin, J.-M. Fortin, F. Pomerleau, and P. Gigu`ere, “Tree detection and diameter estimation based on deep learning,” Forestry: An International Journal of Forest Research , vol. 96, no. 2, pp. 264– 276, 2023

    2023

  21. [29]

    The Auto Arborist Dataset: A Large-Scale Benchmark for Mul- tiview Urban Forest Monitoring Under Domain Shift,

    S. Beery, G. Wu, T. Edwards, et al. , “The Auto Arborist Dataset: A Large-Scale Benchmark for Mul- tiview Urban Forest Monitoring Under Domain Shift,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 21 262–21 275

    2022

  22. [30]

    Tree Species Identification from Bark Images Using Convo- lutional Neural Networks,

    M. Carpentier, P. Gigu `ere, and J. Gaudreault, “Tree Species Identification from Bark Images Using Convo- lutional Neural Networks,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , 2018, pp. 1075–1081

    2018

  23. [31]

    CentralBark Image Dataset and Tree Species Classification Using Deep Learning,

    C. Warner, F. Wu, R. Gazo, B. Benes, N. Kong, and S. Fei, “CentralBark Image Dataset and Tree Species Classification Using Deep Learning,” Algo- rithms, vol. 17, no. 5, p. 179, 2024

    2024

  24. [32]

    Urban street tree dataset for image classification and instance segmentation,

    T. Yang, S. Zhou, Z. Huang, A. Xu, J. Ye, and J. Yin, “Urban street tree dataset for image classification and instance segmentation,” Computers and Electronics in Agriculture, vol. 209, p. 107 852, 2023

    2023

  25. [33]

    The spatial and temporal dynamics of species interactions in mixed-species forests: From pattern to process,

    D. I. Forrester, “The spatial and temporal dynamics of species interactions in mixed-species forests: From pattern to process,” Forest Ecology and Management, vol. 312, pp. 282–292, 2014

    2014

  26. [34]

    Canadian Forest Service, Canada’s National Forest Inventory tree species list , 2014

    2014

  27. [35]

    Exposing the Unseen: Exposure Time Emulation for Offline Benchmarking of Vision Algorithms,

    O. Gamache, J.-M. Fortin, M. Boxan, M. Vaidis, F. Pomerleau, and P. Gigu `ere, “Exposing the Unseen: Exposure Time Emulation for Offline Benchmarking of Vision Algorithms,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , 2024, pp. 11 110–11 117

    2024

  28. [36]

    Swin Transformer: Hier- archical Vision Transformer Using Shifted Windows,

    Z. Liu, Y . Lin, Y . Cao, et al., “Swin Transformer: Hier- archical Vision Transformer Using Shifted Windows,” in IEEE/CVF International Conference on Computer Vision (ICCV), 2021, pp. 10012–10 022

    2021

  29. [37]

    EfficientNetV2: Smaller Models and Faster Training,

    M. Tan and Q. Le, “EfficientNetV2: Smaller Models and Faster Training,” 2021

    2021

  30. [38]

    Minist `ere des Ressources naturelles et des For ˆets, Zones de v ´eg´etation et domaines bioclimatiques du Qu´ebec, 2022

    2022

  31. [39]

    Tree bark re- identification using a deep-learning feature descriptor,

    M. Robert, P. Dallaire, and P. Gigu `ere, “Tree bark re- identification using a deep-learning feature descriptor,” in 17th Conference on Computer and Robot Vision (CRV), 2020, pp. 25–32

    2020

  32. [40]

    Canadian Forest Service, Canada’s National Forest Inventory ground sampling guidelines: specifications for ongoing measurement , 2008

    2008

  33. [41]

    Enhancing Semantic Forestry Segmen- tation Through Advanced Preprocessing With ML Models,

    K. Wołk, J. Niklewski, M. Kopczy ´nski, M. S. Tatara, and O. ˙Zero, “Enhancing Semantic Forestry Segmen- tation Through Advanced Preprocessing With ML Models,” IEEE Access , vol. 13, pp. 98 602–98 621, 2025

    2025

  34. [42]

    Gyawali, M

    A. Gyawali, M. Aalto, and T. Ranta, “Tree Species Detection and Enhancing Semantic Segmentation Us- ing Machine Learning Models with Integrated Multi- spectral Channels from PlanetScope and Digital Aerial Photogrammetry in Young Boreal Forest,” Remote Sensing, vol. 17, no. 11,...

    2025

  35. [43]

    YOLOv12: Attention-Centric Real-Time Object Detectors,

    Y . Tian, Q. Ye, and D. Doermann, “YOLOv12: Attention-Centric Real-Time Object Detectors,” arXiv preprint arXiv:2502.12524, 2025

    2025 arXiv

  36. [44]

    Masked-Attention Mask Transformer for Universal Image Segmentation,

    B. Cheng, I. Misra, A. G. Schwing, A. Kirillov, and R. Girdhar, “Masked-Attention Mask Transformer for Universal Image Segmentation,” in IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 1290–1299

    2022

  37. [45]

    Microsoft COCO: Common Objects in Context,

    T.-Y . Lin, M. Maire, S. Belongie, et al. , “Microsoft COCO: Common Objects in Context,” in European Conference on Computer Vision (ECCV) , Springer, 2014, pp. 740–755

    2014

  38. [46]

    Focal Loss for Dense Ob- ject Detection,

    Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Dollar, “Focal Loss for Dense Ob- ject Detection,” in IEEE International Conference on Computer Vision (ICCV) , 2017, pp. 2999–3007

    2017

  39. [47]

    Using Citizen Science Data as Pre-Training for Se- mantic Segmentation of High-Resolution UA V Images for Natural Forests Post-Disturbance Assessment,

    K. Nasiri, W. Guimont-Martin, D. LaRocque, et al. , “Using Citizen Science Data as Pre-Training for Se- mantic Segmentation of High-Resolution UA V Images for Natural Forests Post-Disturbance Assessment,” Forests, vol. 16, no. 4, p. 616, 2025

    2025

  40. [48]

    Efficient High-Resolution Deep Learning: A Survey,

    A. Bakhtiarnia, Q. Zhang, and A. Iosifidis, “Efficient High-Resolution Deep Learning: A Survey,” ACM Computing Surveys, vol. 56, no. 7, pp. 1–35, 2024

    2024

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.