Exploiting Boundary Loss for the Hierarchical Panoptic Segmentation of Plants and Leaves

Elizabeth Sklar; Madeleine Darbyshire; Simon Parsons

REVIEW 3 major objections 5 minor 30 references

Exploiting Boundary Loss for the Hierarchical Panoptic Segmentation of Plants and Leaves

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A hierarchical panoptic segmentation model for sugar-beet fields improves segmentation of leaves and weeds—and with it leaf-count accuracy—by adding focal and boundary losses to Mask2Former, reaching $PQ^\dagger$ 81.89 on PhenoBench.

desk verdict The title is more honest than the abstract: boundary loss gives a small real gain, but focal loss is a dead end in the ablation as reported. read the letter →

arxiv 2501.00527 v1 pith:DI6DLFHF submitted 2024-12-31 cs.CV cs.LG

Madeleine Darbyshire , Elizabeth Sklar , Simon Parsons This is my paper · ORCID

classification cs.CVcs.LG

keywords hierarchicalpanopticsegmentationprecisionagricultureleafcountingboundarylossfocalMask2FormerPhenoBenchcropandweed

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

This paper tries to show that a hierarchical panoptic segmentation model for farm images can be made better at the small, easy-to-miss objects—individual leaves and weeds—without slowing it down or enlarging it. Starting from Mask2Former, a transformer-based segmentation architecture, the authors add a second transformer decoder so plants and leaves are segmented at the same time, then swap cross-entropy for focal loss and add a boundary loss originally designed for highly unbalanced medical images. On the PhenoBench sugar-beet dataset, the ResNet-50 version improves the combined panoptic metric $PQ^\dagger$ from 81.48 to 81.89 and lowers leaf-counting error (RMSE over ground-truth plants from 2.35 to 2.30). The underlying motivation is practical: better leaf and weed masks mean more accurate leaf counts for growth monitoring and more targeted herbicide application.

What carries the argument

The load-bearing mechanism is the loss function on binary masks. Focal loss ($L_{\text{focal}} = -\alpha_t (1-p_t)^\gamma \log(p_t)$, with $\gamma=2.0$, $\alpha=0.25$) down-weights easy background pixels so that scarce foreground pixels from leaves and weeds contribute more. The boundary loss of Kervadec et al. replaces region integrals with an integral over the interface between the predicted and ground-truth regions, using a precomputed level-set distance map $\phi_G$; its weight starts at 0.01 and grows by 0.0006 per epoch. These losses are applied to masks produced by two separate transformer decoders—one for plants, one for leaves—that share the same learnable queries, with deep supervision at every transformer layer. The loss recipe is what carries the argument: it is what the ablation varies, and it changes segmentation quality at no cost to inference speed.

What would settle it

Retrain the ResNet-50 base and the ResNet-50 focal-plus-boundary variant several times with different random seeds on the standard 1407-image split and compare the distributions of $PQ^\dagger$ and leaf RMSE; if the intervals overlap or the ordering flips, the claimed benefit is not reproducible. A simpler check is whether the same recipe changes $PQ^\dagger$ on the held-out test set in the same direction.

Watch

Extended reading notes

Core claim

The paper's central claim is that on the PhenoBench benchmark, the combination of focal loss and boundary loss, added to an adapted Mask2Former with a separate transformer decoder for leaves, yields better segmentation of small instances and better leaf counts than the unmodified architecture. In the ablation on the standard 1407-image training set with a ResNet-50 backbone, adding both losses raises $PQ^\dagger$ from 81.48 (base) to 81.81, and to 81.89 with test-time augmentation on leaf masks, while the focal-loss-only variant stays essentially flat at 81.46. Leaf-counting error measured as RMSE over all ground-truth plants falls from 2.35 to 2.30 with both losses. With a SwinL backbone the same recipe reaches $PQ^\dagger$ of 83.62 without and 83.70 with test-time augmentation, and it outperforms the published second-place competition submission (81.33) on the same standard split, although it trails the first-place entry that trained on additional validation images.

Load-bearing premise

The reported improvements rest on the assumption that differences of a few tenths of a point in $PQ^\dagger$ and 0.05 in leaf RMSE, measured from single training runs without error bars, reflect the loss functions rather than random variation.

Editorial extensions

If this is right

A lightweight ResNet-50 model can reach panoptic quality competitive with much heavier published systems once focal and boundary losses are added, so small-instance accuracy need not require a large model.
Leaf-count error drops with improved leaf segmentation: RMSE over ground-truth plants falls from 2.35 to 2.30, and over predicted plants from 2.11 to 2.05 for the ResNet-50 model with both losses.
Separating plant and leaf decoders lets one forward pass output both plant-level and leaf-level masks, and the leaf branch can be improved with test-time augmentation without hurting the plant branch.
Because the loss changes do not affect inference speed, the same recipe can be applied to faster architectures for field deployment.

Reading between the lines

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

I infer that the reported $PQ^\dagger$ and leaf-RMSE differences, each below about half a point, should be checked against run-to-run variance; single training runs are not enough to guarantee the focal-plus-boundary recipe is the cause of the improvement.
I infer the boundary-loss idea, designed for highly unbalanced medical segmentations, is likely to transfer to other agricultural or remote-sensing tasks where small objects like weeds, fruits, or seedlings dominate the error.
I infer that leaf-count accuracy may need a counting-specific head or post-processing, since even the paper's best segmentation leaves a nontrivial RMSE and segmentation gains translate into only modest counting gains.
I infer that combining the loss recipe with a stronger backbone (for instance, the BEiT backbone used by the first-place entry) might close the remaining gap to 82.62, but this is an extrapolation from the paper's backbone comparison.

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.

Desk Editor's Note

read the letter

Quick take: this paper is a useful data point on boundary loss, and a cautionary example of overclaiming a combination. The title says boundary loss; the abstract credits both focal and boundary loss. The ablation table only supports the boundary half.

What's new: a hierarchical Mask2Former with a separate leaf decoder, applied to PhenoBench, plus focal loss and boundary loss. That specific package hasn't been tested on this benchmark before. The paper also reports leaf-counting RMSE, which is the metric that matters for the intended application. Code is on GitHub, so the work is reproducible.

What it does well: the ablation is clean to read. Boundary loss alone raises PQ† by 0.30 over the base ResNet-50 model (81.48 -> 81.78), and weed PQ jumps from 49.17 to 51.16. Leaf RMSE_GT drops from 2.35 to 2.31. The authors are candid that leaf-counting gains are modest and that PQ improvements don't always reduce counting error. That honesty earns credit.

The soft spots are exactly where the stress-test note lands. Focal loss alone is a wash: 81.46 vs. 81.48 base. Adding focal loss on top of boundary loss changes PQ† by 0.03 and RMSE_GT by 0.01. That is not a measurable effect on single runs. The conclusion that the combination improves small-instance segmentation is therefore not established. The focal loss should either be backed by multi-seed variance or dropped from the claim. Relatedly, there are no error bars anywhere; TTA is applied only to leaves after the authors saw it help there; and the loss weights and boundary alpha schedule are hand-set with no explicit note about validation-set tuning.

Bottom line: the boundary-loss result is real, modest, and worth having in the literature. The focal-loss claim should be reworded or supported. A serious referee should see this—the benchmark comparison and code are useful to anyone working on plant/leaf panoptic segmentation. I'd ask for a few seeds or a corrected claim before acceptance.

Referee Report

3 major / 5 minor

Summary. The paper proposes an adapted Mask2Former for hierarchical panoptic segmentation on the PhenoBench dataset: a second transformer decoder is added to produce leaf-level masks alongside plant-level masks, and the mask loss is changed to a combination of focal loss, Dice loss, and Kervadec et al.'s boundary loss. The authors report ResNet-50 ablations on the validation set, a SwinL model compared with competition baselines on the private test set, and leaf-counting RMSE metrics. They claim that the proposed losses improve segmentation of small instances (leaves and weeds) and leaf-count accuracy without affecting inference speed or model footprint.

Significance. If the claims were fully supported, the contribution would be practically useful: a drop-in loss modification that improves small-instance segmentation on a standard agricultural benchmark, with released code and no runtime cost. The validation-set ablations are single training runs, however, and the reported differences—especially the incremental effect of focal loss on top of boundary loss—are small enough to be explained by seed variation. The paper has useful engineering content, but the central claim that the focal-plus-boundary combination drives the improvement needs stronger empirical support before it can be accepted as stated.

major comments (3)

[Section 5, Tables 1 and 3] The central claim that the combination of focal loss and boundary loss improves leaf and weed segmentation is not established by the reported single-run ablations. Focal-only (ResNet50 (Lf)) gives PQ+ 81.46, slightly below the base 81.48, while boundary-only gives 81.78 and Lf+Lb gives 81.81; the incremental effect of focal loss over boundary loss is 0.03 PQ+ on the validation set and 0.01 RMSE_GT (2.31 to 2.30). The larger sub-metric gains (PQ_leaf 72.06 to 72.29; PQ_weed 51.16 to 51.56) are also on the order of a few tenths of a point without repeated runs. These numbers support boundary loss as the active ingredient, not the focal-plus-boundary combination advertised in the abstract and conclusion. Please report repeated-run means and variance, or reword the contribution to credit boundary loss alone.
[Section 4.3 and Tables 1-2] Test-time augmentation is applied selectively after inspecting validation-set results: the authors state that they tried TTA but 'only found any benefit when we consider the average results on leaf segmentation.' Because this selection is made on the same validation set used for the reported numbers, the TTA rows in Tables 1 and 2 (e.g., PQ_leaf 72.29 to 72.61 and PQ+ 81.81 to 81.89 for ResNet-50) are optimistic. The paper should either report all TTA configurations tried, treat the validation TTA results as a hypothesis to be confirmed on the private test set, or remove TTA from the headline claim.
[Section 5, Table 2] All SwinL results use the proposed Lf+Lb loss, but no SwinL baseline or SwinL boundary-only ablation is provided. Therefore the paper does not demonstrate that the loss changes, rather than the stronger backbone or the separate leaf decoder, are responsible for the SwinL performance (83.62/83.70 on validation, 81.89 on the test set). Since the state-of-the-art comparison in Table 4 is made with this SwinL model, at least one SwinL control is needed to support the attribution of the gains to the proposed losses.

minor comments (5)

[Abstract and Table 4] The phrase 'achieving a PQ+ of 81.89 on the standard training set' is misleading: the result is obtained on the private test set by a model trained on the standard 1407-image training set. Please rephrase to 'trained on the standard training set and evaluated on the PhenoBench test set.'
[Section 3.3] In the definition of RMSE_GT, the sentence 'This shows the counting error accounting for the role of false positive crop predictions' should say 'false negative crop predictions,' because unmatched ground-truth crops (with predicted leaf count zero) are missed detections, not false positives.
[Table 3] The metric names in the table header are corrupted in the manuscript (e.g., 'RM SET P', 'RM SEP red', 'RM SEGT'). They should be formatted as RMSE_TP, RMSE_Pred, and RMSE_GT.
[Section 4.3] Please clarify what 'TTA ... on the leaf segmentation' means operationally: is the leaf mask averaged over all augmentations while plant and weed outputs are computed from a single pass, and are the same augmentations used for the SwinL test-set results in Table 4?
[Section 2] The reference to previous work should read 'Aich et al.' rather than 'Aich et el.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported PQ+ and leaf-count results are measured on a fixed validation set and are not equivalent, by construction, to any fitted input or self-cited prior result.

full rationale

The paper's central claim is empirical: adapting Mask2Former with an extra leaf decoder and adding focal loss plus boundary loss improves small-instance segmentation and leaf-count accuracy on PhenoBench. The loss functions are standard external formulations (focal loss from Lin et al., boundary loss from Kervadec et al.), and the architecture is built on Mask2Former, with no derivation in which the target metric is encoded into the loss or defined in terms of the reported output. The reported PQ, PQ+, and RMSE values are computed on validation data independent of the training objective; the loss weights and hyperparameters are hand-chosen rather than fitted to the evaluation metric. Comparisons to prior work use external published results on the same benchmark, so the evaluation is not self-referential. The only notable weakness is that ablation differences on single training runs are small and no variance estimates are reported, but that is an evidence-quality and statistical-significance concern, not circularity. No load-bearing self-citation chain or fitted-input-renamed-as-prediction step is present.

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

No new objects, forces, or entities are introduced. The contribution is a combination of existing loss functions and an architectural adaptation. The free parameters are loss weights and schedules that are manually chosen, and the axioms are standard transfer-learning and annotation-quality assumptions.

free parameters (6)

lambda_mask = 2.5
Weight for the combined mask loss (focal + dice + boundary) in the total loss, chosen by hand.
lambda_cls = 1.0
Weight for the class loss in the total loss, chosen by hand.
boundary_alpha_initial = 0.01
Initial weight for the boundary loss, increased by 0.0006 per epoch; the schedule is manually set.
boundary_alpha_increment = 0.0006 per epoch
Per-epoch increase of the boundary loss weight, chosen by hand.
focal_gamma = 2.0
Focusing parameter for focal loss, taken from the original focal loss paper rather than tuned here.
focal_alpha = 0.25
Balancing factor for focal loss, taken from the original focal loss paper rather than tuned here.

assumptions (4)

domain assumption The Mask2Former architecture, including masked attention and point sampling, is effective for panoptic segmentation.
The paper builds directly on Cheng et al. [7] and cites its state-of-the-art performance without re-deriving the architecture's effectiveness.
standard math The boundary loss approximation of Kervadec et al. is valid for the small, highly unbalanced leaf and weed masks.
Section 3.2 uses the integral approximation Dist(∂G,∂S) ≈ 2∫_ΔS D_G(q)dq from Kervadec et al. without adapting it to the specific object sizes in this domain.
domain assumption PhenoBench ground-truth annotations, especially leaf instances, are accurate enough that leaf counts derived from masks are meaningful.
The leaf-counting metrics in Section 3.3 assume each ground-truth leaf mask corresponds to one true leaf; annotation errors would directly affect the claimed counting improvements.
domain assumption COCO-pretrained weights provide a useful initialization for agricultural imagery.
Section 4.2 states the model was pretrained on COCO and fine-tuned on PhenoBench; the transfer is assumed to help rather than hurt.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Exploiting Boundary Loss for the Hierarchical Panoptic Segmentation of Plants and Leaves." pith.science (2026). https://pith.science/paper/DI6DLFHF

@misc{pith2026250100527,
  author       = {Pith},
  title        = {Pith review of: Exploiting Boundary Loss for the Hierarchical Panoptic Segmentation of Plants and Leaves},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DI6DLFHF}},
  note         = {Machine review of arXiv:2501.00527}
}

read the original abstract

Precision agriculture leverages data and machine learning so that farmers can monitor their crops and target interventions precisely. This enables the precision application of herbicide only to weeds, or the precision application of fertilizer only to undernourished crops, rather than to the entire field. The approach promises to maximize yields while minimizing resource use and harm to the surrounding environment. To this end, we propose a hierarchical panoptic segmentation method that simultaneously determines leaf count (as an identifier of plant growth)and locates weeds within an image. In particular, our approach aims to improve the segmentation of smaller instances like the leaves and weeds by incorporating focal loss and boundary loss. Not only does this result in competitive performance, achieving a PQ+ of 81.89 on the standard training set, but we also demonstrate we can improve leaf-counting accuracy with our method. The code is available at https://github.com/madeleinedarbyshire/HierarchicalMask2Former.

Figures

Figures reproduced from arXiv: 2501.00527 by the authors.

**Figure 1.** Example images from the PhenoBench dataset. On the top are the full size images and the bottom shows a close-up of the same image. Images (a) are the originals, (b) show whole plants segmented at the pixel level, and (c) show individual leaves segmented at the pixel level. The paper aims to combine crop and weed segmentation as well as leaf segmentation in a single hierarchical panoptic segmentation architecture. T… view at source ↗

**Figure 2.** Adapted Mask2Former architecture with an additional transformer decoder for segmenting leaves. over the regions. This proposes a potential solution to the imbalance we want to address. In this work, we demonstrate how implementing the focal loss [15] and boundary loss [12] can improve segmentation performance. Importantly, for the realworld application of this technology, these can be employed to improve relativel… view at source ↗

**Figure 3.** The boundary of ground truth area G, ∂G, is shown in green and the boundary of segmentation mask S, ∂S is shown in orange. L_{boundary} = \alpha \int _{\Omega } \phi _G(q) s(q) dq (5) where α is a weight that increases throughout training. We initially set α = 0.01 and increase it by 0.0006 at each epoch. s(q) is a function that returns the probability p for each position q. Loss Function While in the original Mask2… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Comparison of ground truth segmentations (top) with the segmentation results on SwinL (Lf + Lb) + TTA (bottom). The grey square in 4b demarcates the zoomed area shown in 4c. Leaf Counting The focus on improving leaf segmentation is further supported when evaluating lea…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 18 canonical work pages

[1]

https://cvppa2023.github

Phenobench: Hierarchical panoptic segmentation. https://cvppa2023.github. io/challenges/#hierarchical- panoptic- segmentation- of- crops- and- weeds (2023)

work page 2023
[2]

In: Proceedings of the IEEE international conference on computer vision workshops

Aich, S., Stavness, I.: Leaf counting with deep convolutional and deconvolutional networks. In: Proceedings of the IEEE international conference on computer vision workshops. pp. 2080–2089 (2017)

work page 2017
[3]

IEEE transactions on pat- tern analysis and machine intelligence39(12), 2481–2495 (2017)

Badrinarayanan, V., Kendall, A., Cipolla, R.: Segnet: A deep convolutional encoder-decoder architecture for image segmentation. IEEE transactions on pat- tern analysis and machine intelligence39(12), 2481–2495 (2017)

2017
[4]

arXiv preprint arXiv:2106.08254 (2021)

Bao, H., Dong, L., Piao, S., Wei, F.: Beit: Bert pre-training of image transformers. arXiv preprint arXiv:2106.08254 (2021)

arXiv 2021
[5]

arXiv preprint arXiv:1706.05587 (2017)

Chen, L.C., Papandreou, G., Schroff, F., Adam, H.: Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587 (2017)

arXiv 2017
[6]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Cheng, B., Collins, M.D., Zhu, Y., Liu, T., Huang, T.S., Adam, H., Chen, L.C.: Panoptic-deeplab: A simple, strong, and fast baseline for bottom-up panoptic seg- mentation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 12475–12485 (2020)

2020
[7]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Cheng, B., Misra, I., Schwing, A.G., Kirillov, A., Girdhar, R.: Masked-attention mask transformer for universal image segmentation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 1290–1299 (2022)

2022
[8]

Advances in Neural Information Processing Systems34, 17864–17875 (2021)

Cheng, B., Schwing, A., Kirillov, A.: Per-pixel classification is not all you need for semantic segmentation. Advances in Neural Information Processing Systems34, 17864–17875 (2021)

2021

Show all 30 references

[9]

In: 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Di Cicco, M., Potena, C., Grisetti, G., Pretto, A.: Automatic model based dataset generation for fast and accurate crop and weeds detection. In: 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 5188–

2017
[10]

Computers and Electronics in Agriculture202, 107388 (2022)

Genze, N., Ajekwe, R., Güreli, Z., Haselbeck, F., Grieb, M., Grimm, D.G.: Deep learning-based early weed segmentation using motion blurred UAV images of sorghum fields. Computers and Electronics in Agriculture202, 107388 (2022)

2022
[11]

He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition (2015)

2015
[12]

In: International conference on medical imaging with deep learning

Kervadec, H., Bouchtiba, J., Desrosiers, C., Granger, E., Dolz, J., Ayed, I.B.: Boundary loss for highly unbalanced segmentation. In: International conference on medical imaging with deep learning. pp. 285–296. PMLR (2019) 14 M. Darbyshire et al

2019
[13]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition

Kirillov, A., He, K., Girshick, R., Rother, C., Dollár, P.: Panoptic segmentation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition. pp. 9404–9413 (2019)

2019
[14]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision

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

2023
[15]

Lin, T.Y., Goyal, P., Girshick, R., He, K., Dollár, P.: Focal loss for dense object detection.In:ProceedingsoftheIEEEinternationalconferenceoncomputervision. pp. 2980–2988 (2017)

2017
[16]

In:Proceedings of the IEEE/CVF international conference on computer vision

Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin transformer:Hierarchicalvision transformer usingshiftedwindows. In:Proceedings of the IEEE/CVF international conference on computer vision. pp. 10012–10022 (2021)

2021
[17]

arXiv preprint arXiv:1711.05101 (2017)

Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

2017 arXiv
[18]

Medical Image Analysis71, 102035 (2021)

Ma, J., Chen, J., Ng, M., Huang, R., Li, Y., Li, C., Yang, X., Martel, A.L.: Loss odyssey in medical image segmentation. Medical Image Analysis71, 102035 (2021)

2021
[19]

AgriEngi- neering 2(3), 471–488 (2020)

Osorio, K., Puerto, A., Pedraza, C., Jamaica, D., Rodríguez, L.: A deep learning approach for weed detection in lettuce crops using multispectral images. AgriEngi- neering 2(3), 471–488 (2020)

2020
[20]

In: 2023 IEEE International Conference on Robotics and Automation (ICRA)

Roggiolani, G., Sodano, M., Guadagnino, T., Magistri, F., Behley, J., Stachniss, C.: Hierarchical approach for joint semantic, plant instance, and leaf instance seg- mentation in the agricultural domain. In: 2023 IEEE International Conference on Robotics and Automation (ICRA)....

2023
[21]

IEEE Transactions on In- telligent Transportation Systems19(1), 263–272 (2017)

Romera, E., Alvarez, J.M., Bergasa, L.M., Arroyo, R.: Erfnet: Efficient residual factorized convnet for real-time semantic segmentation. IEEE Transactions on In- telligent Transportation Systems19(1), 263–272 (2017)

2017
[22]

In: Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, Oc- tober 5-9, 2015, Proceedings, Part III 18

Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomed- ical image segmentation. In: Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, Oc- tober 5-9, 2015, Proceedings, Part III 18....

2015
[23]

IEEE Robotics and Automation Letters 3(1), 588–595 (Jan 2018)

Sa, I., Chen, Z., Popović, M., Khanna, R., Liebisch, F., Nieto, J., Siegwart, R.: weedNet: Dense semantic weed classification using multispectral images and MAV for smart farming. IEEE Robotics and Automation Letters 3(1), 588–595 (Jan 2018). https://doi.org/10.1109/LRA.2017.2774979

2018
[24]

In: European conference on computer vision

Scharr, H., Minervini, M., Fischbach, A., Tsaftaris, S.A.: Annotated image datasets of rosette plants. In: European conference on computer vision. Zürich, Suisse. pp. 6–12 (2014)

2014
[25]

arXiv preprint (2023)

Weyler, J., Magistri, F., Marks, E., Chong, Y.L., Sodano, M., Roggiolani, G., Chebrolu, N., Stachniss, C., Behley, J.: PhenoBench — A Large Dataset and Benchmarks for Semantic Image Interpretation in the Agricultural Domain. arXiv preprint (2023)

2023
[26]

In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision

Weyler, J., Magistri, F., Seitz, P., Behley, J., Stachniss, C.: In-field phenotyp- ing based on crop leaf and plant instance segmentation. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. pp. 2725– 2734 (2022)

2022
[27]

IEEE Robotics and Automation Letters 7(2), 3787–3794 (2022) Boundary Loss for Hierarchical Panoptic Segmentation 15

Weyler, J., Quakernack, J., Lottes, P., Behley, J., Stachniss, C.: Joint plant and leaf instance segmentation on field-scale UAV imagery. IEEE Robotics and Automation Letters 7(2), 3787–3794 (2022) Boundary Loss for Hierarchical Panoptic Segmentation 15

2022
[28]

Wu, Y., Kirillov, A., Massa, F., Lo, W.Y., Girshick, R.: Detectron2.https:// github.com/facebookresearch/detectron2 (2019)

2019
[29]

arXiv preprint arXiv:2010.04159 (2020)

Zhu, X., Su, W., Lu, L., Li, B., Wang, X., Dai, J.: Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159 (2020)

2020 arXiv
[30]

Computers and Electronics in Agriculture187, 106242 (2021)

Zou, K., Chen, X., Wang, Y., Zhang, C., Zhang, F.: A modified U-Net with a specific data argumentation method for semantic segmentation of weed images in the field. Computers and Electronics in Agriculture187, 106242 (2021)

2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

https://cvppa2023.github

Phenobench: Hierarchical panoptic segmentation. https://cvppa2023.github. io/challenges/#hierarchical- panoptic- segmentation- of- crops- and- weeds (2023)

work page 2023

[2] [2]

In: Proceedings of the IEEE international conference on computer vision workshops

Aich, S., Stavness, I.: Leaf counting with deep convolutional and deconvolutional networks. In: Proceedings of the IEEE international conference on computer vision workshops. pp. 2080–2089 (2017)

work page 2017

[3] [3]

IEEE transactions on pat- tern analysis and machine intelligence39(12), 2481–2495 (2017)

Badrinarayanan, V., Kendall, A., Cipolla, R.: Segnet: A deep convolutional encoder-decoder architecture for image segmentation. IEEE transactions on pat- tern analysis and machine intelligence39(12), 2481–2495 (2017)

2017

[4] [4]

arXiv preprint arXiv:2106.08254 (2021)

Bao, H., Dong, L., Piao, S., Wei, F.: Beit: Bert pre-training of image transformers. arXiv preprint arXiv:2106.08254 (2021)

arXiv 2021

[5] [5]

arXiv preprint arXiv:1706.05587 (2017)

Chen, L.C., Papandreou, G., Schroff, F., Adam, H.: Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587 (2017)

arXiv 2017

[6] [6]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Cheng, B., Collins, M.D., Zhu, Y., Liu, T., Huang, T.S., Adam, H., Chen, L.C.: Panoptic-deeplab: A simple, strong, and fast baseline for bottom-up panoptic seg- mentation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 12475–12485 (2020)

2020

[7] [7]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Cheng, B., Misra, I., Schwing, A.G., Kirillov, A., Girdhar, R.: Masked-attention mask transformer for universal image segmentation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 1290–1299 (2022)

2022

[8] [8]

Advances in Neural Information Processing Systems34, 17864–17875 (2021)

Cheng, B., Schwing, A., Kirillov, A.: Per-pixel classification is not all you need for semantic segmentation. Advances in Neural Information Processing Systems34, 17864–17875 (2021)

2021