Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

RMT-PPAD: Real-time Multi-task Learning for Panoptic Perception in Autonomous Driving

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read RMT-PPAD claims one real-time transformer network beats all open-source multi-task baselines on BDD100K detection, drivable area, and lane lines at 32.6 FPS.

desk verdict Solid MTL engineering with a genuinely useful evaluation fix; the lane-label SOTA number is conditional on a plausible but unverified premise. read the letter →

arxiv 2508.06529 v1 pith:BMQG2OL3 submitted 2025-08-02 cs.CV cs.LG

classification cs.CVcs.LG
keywords multi-tasklearningpanopticdrivingperceptionobjectdetectiondrivableareasegmentationlanelinenegativetransferreal-timetransformerBDD100K
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

RMT-PPAD is a real-time, transformer-based multi-task network that performs object detection, drivable-area segmentation, and lane-line segmentation in a single forward pass. The paper claims that on BDD100K it beats every open-source multi-task driving-perception model it compares with on all three tasks: mAP50 of 84.9%, Recall of 95.4%, drivable-area mIoU of 92.6%, and lane-line IoU of 56.8% with accuracy of 84.7%, while running at 32.6 FPS. The central engineering claim is that a lightweight gate-control-with-adapter (GCA) module fuses shared and task-specific features so that the tasks help rather than hurt one another. The paper also claims that lane-line evaluations in earlier work were unfair because models were trained on 8-pixel-wide lane labels but tested against 2-pixel-wide labels, and that dilating the test labels to 8 pixels gives a fairer comparison. If the paper is right, real-time multi-task panoptic perception can be achieved without manually designed task-specific heads, and a simple label-alignment fix can change published rankings.

What carries the argument

The central object is the gate control with adapter (GCA), a lightweight module inserted at three scales between the efficient hybrid encoder and the task decoders. For each scale, a task adapter, built from $1\times1$ convolutions and a depthwise separable convolution, extracts task-specific features while leaving shared features intact; a gate then applies squeeze-and-excitation-style channel attention, spatial attention, and a fusion gate module, clipping the combined gate to $[0.05, 0.95]$ so no stream is fully suppressed. The fusion rule is $out = shared + gate \cdot (task - shared)$, a residual interpolation that lets each task emphasize useful shared features and suppress conflicting ones. A second mechanism is the adaptive segmentation decoder, which stacks the three scale feature maps and multiplies them by softmax-normalized learnable weights $\alpha \in \mathbb{R}^{2\times3}$, one row per segmentation task, avoiding manually designed heads for drivable area versus lane lines. The third mechanism is the label-alignment fix: dilating lane-line test labels to the 8-pixel width used in training, which the paper argues makes lane-line IoU a faithful measure.

What would settle it

Download the lane-line annotation files and the training and evaluation scripts of YOLOP, HybridNet, YOLOPX, and A-YOLOM, and measure the actual pixel-width distributions of lane masks in training versus test. If the training labels are not predominantly 8 pixels wide or the test labels are not predominantly 2 pixels wide, the paper's unfair-evaluation diagnosis and the resulting lane-IoU rankings in Table III collapse. Alternatively, re-evaluate all models on the original 2-pixel test labels; if RMT-PPAD does not still lead in lane IoU, the label-dilation fix, not the model, is carrying the lane result.

Watch

Extended reading notes

Core claim

The paper's claim is that a single RT-DETR-based network with six lightweight GCA modules can jointly predict vehicles, drivable area, and lane lines better than all open-source multi-task baselines on BDD100K while still running above the 30 FPS camera sampling rate. The load-bearing novelty is the GCA module: an adapter produces task-specific features from shared multi-scale features, and a dynamic gate combining channel attention, spatial attention, and a fusion gate computes $out = shared + gate \cdot (task - shared)$ with the gate clipped to $[0.05, 0.95]$, which the authors say alleviates negative transfer; their gradient-cosine-similarity analysis supports this by showing fewer conflicting gradients in the three task pairs. The paper further claims that the previous lane-line protocol was inconsistent: models were trained on 8-pixel-wide lane labels but evaluated against 2-pixel-wide test labels, so more accurate predictions accumulated false positives and were scored worse; dilating the test labels with a $7\times7$ elliptical structuring element restores agreement with visual quality. A unified segmentation decoder with a learnable softmax-normalized weight tensor $\alpha$ over the three scales learns task-appropriate preferences, such as drivable area relying more on high-level features and lane lines relying more on low-level detail.

Load-bearing premise

The results hinge on the claim that earlier multi-task models trained lane lines with 8-pixel-wide labels and were tested against 2-pixel-wide labels, so dilating the test labels to 8 pixels is a fair correction rather than a change of the task.

Editorial extensions

If this is right

  • If the GCA result holds, multi-task driving perception can be built with a shared backbone plus a lightweight gating module and no task-specific segmentation heads, reducing design and engineering cost.
  • If the lane-label dilation correction is accepted, the published lane-line IoU rankings of open-source models such as YOLOP, HybridNet, YOLOPX, and A-YOLOM should be recomputed, since some models are penalized more than others by thick predictions compared with 2-pixel ground truth.
  • The 32.6 FPS result implies the model runs above the 30 FPS camera sampling rate on a single RTX 4090 without inference accelerators, meeting the paper's definition of real-time for autonomous driving.
  • The gradient cosine-similarity evidence indicates that the GCA reduces conflicts among the three tasks, so joint training need not degrade individual task performance.

Reading between the lines

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

  • Our inference: the label-width inconsistency probably affects any benchmark where training and test annotations use different line thicknesses, so the same dilation audit could be applied to other road-marking or lane datasets.
  • Our inference: the GCA module is task-agnostic and could be inserted into other multi-task dense-prediction models; its benefit could be tested by measuring gradient cosine similarity before and after insertion on depth, segmentation, and detection combinations.
  • Our inference: the learned weights $\alpha$ offer a diagnostic for which feature scales each task relies on; monitoring them under domain shift might reveal when a model starts depending on the wrong resolution.
  • Our inference: the real-world dashcam evaluation in the paper is qualitative, so a quantitative extension with manually labelled real-road frames would be needed to confirm the claimed stability in practice.
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. This manuscript introduces RMT-PPAD, a multi-task transformer for panoptic driving perception that jointly handles vehicle detection, drivable-area segmentation, and lane-line segmentation. It proposes a Gate Control with Adapter (GCA) module to fuse shared and task-specific features, an adaptive segmentation decoder that learns per-task multi-scale weighting, and a lane-line evaluation correction in which test masks are dilated from roughly 2 px to 8 px to match the training masks used by prior work. On BDD100K, the authors report mAP50 84.9%, Recall 95.4%, drivable-area mIoU 92.6%, lane-line IoU 56.8% and ACC 84.7% at 32.6 FPS on a single RTX 4090, claiming SOTA over open-source panoptic-driving MTL baselines. Ablations on a 10K-image toy split and gradient-similarity histograms support the negative-transfer claim, and qualitative real-world results are provided.

Significance. The paper has real strengths: the authors release code and weights, re-run all comparison methods under the same dilated-label evaluation protocol (the right way to handle the metric change), and provide a quantitative gradient-similarity analysis rather than only qualitative claims. If the lane-label premise is correct and the toy-set ablations transfer, the GCA module and the adaptive decoder are useful architectural contributions with immediate practical value for real-time autonomous-driving perception. However, the headline lane-line SOTA is tightly coupled to an unverified assumption about every baseline's training-label width, and the main negative-transfer evidence is obtained on a reduced dataset; as presented, the significance is moderate and the strongest claims need additional support.

major comments (3)
  1. [Section III-D, Table III] The headline lane-line results are conditional on the premise that every compared baseline trained with 8-pixel-wide lane masks and was evaluated against 2-pixel masks, so that dilating the test masks to 8 pixels restores a fair comparison. The manuscript does not verify this premise for YOLOP, HybridNet, YOLOPX, A-YOLOM(n), or A-YOLOM(s); it only cites [14] and states the practice in [3]-[5], [11]-[13]. It also does not measure how well the 7x7 elliptical dilation, which the authors concede yields 8- or 9-pixel widths and a systematic error on diagonals in Section III-D, matches the actual training masks of each baseline. Because Fig. 6 shows the protocol change taking YOLOPX IoU from 0.2339 to 0.5851 on one sample, the 4.7-point IoU margin over YOLOPX in Table III could be a metric artifact rather than a model advantage. I ask the authors to (i) state how RMT-PPAD's own training masks are generated, (ii) confirm from released code or documentation the training-label width of each baseline, and (iii) report each method's lane IoU/ACC under both the original 2-pixel protocol and the dilated 8-pixel protocol so the effect of the protocol change is visible per method.
  2. [Section IV-A.3, Section IV-C, Table IV] The central evidence that GCA alleviates negative transfer is produced on a toy dataset of 10K training and 2K validation images, not on the full BDD100K split. The numbers in Table IV (e.g., RMT-PPAD mAP50 78.3% and mIoU 91.3%) are far from the full-model results in Table III (84.9% and 92.6%), so the conclusion that GCA 'even outperforms the single-task baselines in mAP50 and mIoU' and 'fosters synergy across tasks' is established only in a reduced-data regime. The paper's contribution statements in the abstract and Section I do not carry this caveat. The authors should either repeat the single-task/vanilla-MTL/with-GCA comparison on the full training set (or a substantially larger subset with full epochs), or explicitly restrict the negative-transfer claims to the toy setting and explain why the trend is expected to transfer.
  3. [Section IV-A.3, Table V] The segmentation mask thresholds used for the reported BDD100K numbers (0.45 for drivable area, 0.9 for lane lines) are selected on the validation set from Table V, and all results are single-run without error bars or seed variation. This raises two issues: the margins in Table III (for example, 4.7 IoU over HybridNet in lane lines) may be threshold-dependent, and it is not clear whether the baselines were evaluated with their own optimal thresholds or with the same fixed thresholds. The paper should report the sensitivity of each method to the threshold, give multi-seed variance for at least the main comparison, and state the exact evaluation protocol for all baselines.
minor comments (6)
  1. [Section I] Sentences such as "We through experiments and ablation studies to demonstrate..." and "RMT-PPAD achieve outperforming while maintaining real-time performance" contain grammatical errors and need revision.
  2. [Section II-C] The sentence "YOLOP [5] YOLOP introduces..." has a duplicated subject and should be corrected.
  3. [Section IV-A.3, Table III] The notation "A-YOLOM(n)" and "A-YOLOM(s)" is used without defining what (n) and (s) mean; the authors should state that these are the nano and small variants, or otherwise explain the suffix.
  4. [Figure 5] The block labeled "S Deconv" is ambiguous; a full name or a clearer caption would help the reader understand the decoder structure.
  5. [Section IV-D] The text says the real-road dataset "includes four scenarios: snow, cloudy, and daytime," which lists only three scenarios, while Fig. 10 caption lists "nighttime, daytime, cloudy, and highway"; please reconcile the enumeration.
  6. [Section III-D] "Table I(c) and I(e) shows" should be "show," and the phrase "previous work labels in Table I(b) and I(d)" would be clearer if the subfigures were referred to by their explicit labels (a)-(e).

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity; evaluation-protocol correction and non-load-bearing self-citation do not reduce the central claims to their inputs.

full rationale

RMT-PPAD's central claims are empirical and self-contained: the reported mAP50, mIoU, and lane-line IoU/ACC numbers in Table III are obtained by training the network on BDD100K and measuring predictions against ground-truth masks, not by fitting a parameter to the evaluation numbers. The GCA equations (1)-(2), the learnable segmentation weights alpha in Fig. 5, and the loss functions (5)-(17) are trained or manually set hyperparameters; none is derived from the headline metrics. The lane-line label analysis in Section III-D is an evaluation-protocol correction: the paper dilates test labels to 8-pixel width to match the stated 8-pixel training labels and recomputes IoU/ACC for all baselines under the corrected protocol. This changes the metric but does not define the model's outputs, and the 'state-of-the-art' claim is conditional on that protocol rather than circular. The only self-citation is the authors' own A-YOLOM baseline [4], used as an open-source comparison method; it is not load-bearing for the GCA design or the SOTA claim. Whether the 8-pixel training-label premise holds for every baseline is a correctness/protocol risk, not a circularity, so no circular step is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper's claims depend mainly on standard neural training assumptions rather than on a mathematical derivation. The notable free choices are loss weights, mask thresholds (selected on the validation set), and the learned segmentation scale weights. No invented physical entities are introduced.

free parameters (3)
  • Segmentation mask thresholds = 0.45 (drivable area), 0.90 (lane line)
    Selected as best-performing thresholds on BDD100K validation in Table V; final reported numbers use these values.
  • Loss coefficients = alpha=1, beta=5, gamma=2, lambda_fl=24, lambda_bce=8, lambda_tv=8
    Chosen by hand in Section IV-A.3; central results depend on them, with no sensitivity analysis.
  • Learnable segmentation scale weights alpha = drivable [0.355, 0.156, 0.489], lane [0.405, 0.442, 0.153]
    Trained on BDD100K and used to weight multi-scale segmentation features; the paper presents the learned values as evidence of task-specific preferences.
assumptions (4)
  • domain assumption BDD100K validation set is a reliable proxy for real autonomous driving performance.
    All quantitative conclusions in Table III are drawn from a single split of BDD100K; no separate held-out test set is used because test annotations are unavailable.
  • domain assumption Prior panoptic-driving MTL methods trained lane lines with 8-pixel-wide labels and evaluated against 2-pixel-wide labels.
    Used in Section III-D to justify dilating test labels; if false, the lane line comparison and the proposed correction lose their foundation.
  • domain assumption Real-time performance is defined as above 30 FPS on a single RTX 4090 without inference accelerators.
    The paper's real-time claim uses this hardware-specific threshold; performance under deployment constraints on vehicle hardware is not measured.
  • domain assumption Gradient cosine similarity distributions are a valid indicator of negative transfer.
    Section IV-C.2 uses histograms of gradient similarities to claim GCA alleviates task conflicts; no quantitative link between these histograms and final task accuracy is established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RMT-PPAD: Real-time Multi-task Learning for Panoptic Perception in Autonomous Driving." pith.science (2026). https://pith.science/paper/BMQG2OL3

@misc{pith2026250806529,
  author       = {Pith},
  title        = {Pith review of: RMT-PPAD: Real-time Multi-task Learning for Panoptic Perception in Autonomous Driving},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BMQG2OL3}},
  note         = {Machine review of arXiv:2508.06529}
}
read the original abstract

Autonomous driving systems rely on panoptic driving perception that requires both precision and real-time performance. In this work, we propose RMT-PPAD, a real-time, transformer-based multi-task model that jointly performs object detection, drivable area segmentation, and lane line segmentation. We introduce a lightweight module, a gate control with an adapter to adaptively fuse shared and task-specific features, effectively alleviating negative transfer between tasks. Additionally, we design an adaptive segmentation decoder to learn the weights over multi-scale features automatically during the training stage. This avoids the manual design of task-specific structures for different segmentation tasks. We also identify and resolve the inconsistency between training and testing labels in lane line segmentation. This allows fairer evaluation. Experiments on the BDD100K dataset demonstrate that RMT-PPAD achieves state-of-the-art results with mAP50 of 84.9% and Recall of 95.4% for object detection, mIoU of 92.6% for drivable area segmentation, and IoU of 56.8% and accuracy of 84.7% for lane line segmentation. The inference speed reaches 32.6 FPS. Moreover, we introduce real-world scenarios to evaluate RMT-PPAD performance in practice. The results show that RMT-PPAD consistently delivers stable performance. The source codes and pre-trained models are released at https://github.com/JiayuanWang-JW/RMT-PPAD.

Figures

Figures reproduced from arXiv: 2508.06529 by the authors.

Figure 1
Figure 1. Fundamental tasks in panoptic driving perception: [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of RMT-PPAD structure. region is classified independently, and the predicted label is propagated to all pixels within the region. However, this approach has a significant limitation in real-time inference applications because it has a multi-stage pipeline instead of an end-to-end pipeline. To address this limitation, fully convolutional networks [28] (FCN) were proposed. It directly predicts classification … view at source ↗
Figure 3
Figure 3. Overview of GCA structure. DW Conv denotes [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Structure of the detection decoder. Yellow boxes are [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Structure of Segmentation. α is a learnable weight tensor. optimizes object queries to generate bounding boxes (Bboxes) and confidence scores (Scores). Segmentation decoder: We adopt a unified segmentation decoder for two tasks. This design is different from previous s…
Figure 6
Figure 6. Figure 6: Lane line test label compares explanation. (a) is ground truth (GT) for 2-pixel, which is the previous work used label [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Illustration of lane lines. The pink area is a 2-pixel [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Visualization results comparison on different scenarios. From left to right are night, snow, rain, and daytime. [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: The gradients cosine similarity of each task in vanilla MTL and RMT-PPAD. Among them, tasks 1 to 3 represent [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Real road results. From left to right are nighttime, daytime, cloudy, and highway scenarios. [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]

Discussion (0). Continue with ORCID 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

  1. Citation Issues in Wave Mechanics Theory of Microwave Absorption

    physics.soc-ph 2025-08 unverdicted novelty 6.0 of 10

    The paper claims that impedance-matching and quarter-wavelength rules for thin-film microwave absorbers are flawed, that wave interference at the film's surfaces is the correct mechanism, and that citation neglect of ...

Reference graph

Works this paper leans on

35 extracted references · 26 canonical work pages · cited by 1 Pith paper

  1. [14]

    Learning lightweight lane detection cnns by self attention distillation,

    Y . Hou, Z. Ma, C. Liu, and C. C. Loy, “Learning lightweight lane detection cnns by self attention distillation,” in Proceedings of the IEEE/CVF international conference on computer vision , 2019, pp. 1013– 1021

  2. [3]

    Yolopx: Anchor- free multi-task learning network for panoptic driving perception,

    J. Zhan, Y . Luo, C. Guo, Y . Wu, J. Meng, and J. Liu, “Yolopx: Anchor- free multi-task learning network for panoptic driving perception,” Pat- tern Recognition , vol. 148, p. 110152, 2024

  3. [5]

    Yolop: You only look once for panoptic driving perception,

    D. Wu, M.-W. Liao, W.-T. Zhang, X.-G. Wang, X. Bai, W.-Q. Cheng, and W.-Y . Liu, “Yolop: You only look once for panoptic driving perception,” Machine Intelligence Research, vol. 19, no. 6, pp. 550–562, 2022

  4. [11]

    Hybridnets: End-to-end perception network,

    D. Vu, B. Ngo, and H. Phan, “Hybridnets: End-to-end perception network,” arXiv preprint arXiv:2203.09035 , 2022

  5. [13]

    Yolopv2: Better, faster, stronger for panoptic driving perception,

    C. Han, Q. Zhao, S. Zhang, Y . Chen, Z. Zhang, and J. Yuan, “Yolopv2: Better, faster, stronger for panoptic driving perception,” arXiv preprint arXiv:2208.11434, 2022

  6. [1]

    Multi-task learning for dense prediction tasks: A survey,

    S. Vandenhende, S. Georgoulis, W. Van Gansbeke, M. Proesmans, D. Dai, and L. Van Gool, “Multi-task learning for dense prediction tasks: A survey,” IEEE transactions on pattern analysis and machine intelligence, vol. 44, no. 7, pp. 3614–3633, 2021

  7. [2]

    Autonomous driving cars in smart cities: Recent advances, requirements, and challenges,

    I. Yaqoob, L. U. Khan, S. A. Kazmi, M. Imran, N. Guizani, and C. S. Hong, “Autonomous driving cars in smart cities: Recent advances, requirements, and challenges,” IEEE Network , vol. 34, no. 1, pp. 174– 181, 2019

  8. [4]

    You only look at once for real-time and generic multi-task,

    J. Wang, Q. J. Wu, and N. Zhang, “You only look at once for real-time and generic multi-task,” IEEE Transactions on V ehicular Technology , 2024

Show all 35 references
  1. [6]

    A loss-balanced multi-task model for simultaneous detection and segmentation,

    W. Zhang, K. Wang, Y . Wang, L. Yan, and F.-Y . Wang, “A loss-balanced multi-task model for simultaneous detection and segmentation,” Neuro- computing, vol. 428, pp. 65–78, 2021

  2. [7]

    Multi-task learning with attention for end-to-end autonomous driving,

    K. Ishihara, A. Kanervisto, J. Miura, and V . Hautamaki, “Multi-task learning with attention for end-to-end autonomous driving,” in Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 2902–2911

  3. [8]

    t-readi: Transformer-powered robust and efficient multimodal inference for autonomous driving,

    P. Hu, Y . Qian, T. Zheng, A. Li, Z. Chen, Y . Gao, X. Cheng, and J. Luo, “t-readi: Transformer-powered robust and efficient multimodal inference for autonomous driving,” IEEE Transactions on Mobile Computing , 2024

  4. [9]

    Fast quantum convolutional neural networks for low-complexity object detection in autonomous driving applications,

    E. J. Roh, H. Baek, D. Kim, and J. Kim, “Fast quantum convolutional neural networks for low-complexity object detection in autonomous driving applications,” IEEE Transactions on Mobile Computing , 2024

  5. [10]

    Deep cnn-based real-time traffic light detector for self-driving vehicles,

    Z. Ouyang, J. Niu, Y . Liu, and M. Guizani, “Deep cnn-based real-time traffic light detector for self-driving vehicles,” IEEE Transactions on Mobile Computing , vol. 19, no. 2, pp. 300–313, 2019

  6. [12]

    Research on road scene understanding of autonomous vehicles based on multi-task learning,

    J. Guo, J. Wang, H. Wang, B. Xiao, Z. He, and L. Li, “Research on road scene understanding of autonomous vehicles based on multi-task learning,” Sensors, vol. 23, no. 13, p. 6238, 2023

  7. [15]

    Detrs beat yolos on real-time object detection,

    Y . Zhao, W. Lv, S. Xu, J. Wei, G. Wang, Q. Dang, Y . Liu, and J. Chen, “Detrs beat yolos on real-time object detection,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2024, pp. 16 965–16 974

  8. [16]

    Bdd100k: A diverse driving dataset for heterogeneous multitask learning,

    F. Yu, H. Chen, X. Wang, W. Xian, Y . Chen, F. Liu, V . Madhavan, and T. Darrell, “Bdd100k: A diverse driving dataset for heterogeneous multitask learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020, pp. 2636–2645

  9. [17]

    Rich feature hierarchies for accurate object detection and semantic segmentation,

    R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2014, pp. 580–587

  10. [18]

    Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,

    C.-Y . Wang, A. Bochkovskiy, and H.-Y . M. Liao, “Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 7464–7475

  11. [19]

    A robust anchor-free detection method for sar ship targets with lightweight cnn,

    Y . Hao, J. Wu, Y . Yao, and Y . Guo, “A robust anchor-free detection method for sar ship targets with lightweight cnn,” IEEE Transactions on Instrumentation and Measurement , 2025

  12. [20]

    Faa-det: Feature augmentation and alignment for anchor-free oriented object detection,

    Z. Li, W. Liu, Z. Xie, X. Kang, P. Duan, and S. Li, “Faa-det: Feature augmentation and alignment for anchor-free oriented object detection,” IEEE Transactions on Geoscience and Remote Sensing , 2024

  13. [21]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” in European conference on computer vision . Springer, 2020, pp. 213– 229

  14. [22]

    Deformable detr: Deformable transformers for end-to-end object detection,

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

  15. [23]

    Region-based semantic segmen- tation with end-to-end training,

    H. Caesar, J. Uijlings, and V . Ferrari, “Region-based semantic segmen- tation with end-to-end training,” in European Conference on Computer Vision. Springer, 2016, pp. 381–397

  16. [24]

    Region-based convolutional networks for accurate object detection and segmentation,

    R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Region-based convolutional networks for accurate object detection and segmentation,” IEEE transactions on pattern analysis and machine intelligence , vol. 38, no. 1, pp. 142–158, 2015

  17. [25]

    Std2p: Rgbd semantic segmentation using spatio-temporal data-driven pooling,

    Y . He, W.-C. Chiu, M. Keuper, and M. Fritz, “Std2p: Rgbd semantic segmentation using spatio-temporal data-driven pooling,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2017, pp. 4837–4846

  18. [26]

    Per-pixel classification is not all you need for semantic segmentation,

    B. Cheng, A. Schwing, and A. Kirillov, “Per-pixel classification is not all you need for semantic segmentation,” Advances in neural information processing systems, vol. 34, pp. 17 864–17 875, 2021

  19. [27]

    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 Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 1290–1299

  20. [28]

    Fully convolutional networks for semantic segmentation,

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

  21. [29]

    Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,

    L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,” IEEE transactions on pattern analysis and machine intelligence , vol. 40, no. 4, pp. 834–848, 2017

  22. [30]

    Pyramid scene parsing network,

    H. Zhao, J. Shi, X. Qi, X. Wang, and J. Jia, “Pyramid scene parsing network,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2017, pp. 2881–2890

  23. [31]

    Yolomh: You only look once for multi-task driving perception with high efficiency,

    L. Fang, S. Bowen, M. Jianxi, and S. Weixing, “Yolomh: You only look once for multi-task driving perception with high efficiency,” Machine Vision and Applications , vol. 35, no. 3, p. 44, 2024

  24. [32]

    Squeeze-and-excitation networks,

    J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 7132–7141. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, JANUARY 2025 13

  25. [33]

    Generalized intersection over union: A metric and a loss for bound- ing box regression,

    H. Rezatofighi, N. Tsoi, J. Gwak, A. Sadeghian, I. Reid, and S. Savarese, “Generalized intersection over union: A metric and a loss for bound- ing box regression,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern recognition , 2019, pp. 658–666

  26. [34]

    Focal loss for dense object detection,

    T.-Y . Lin, P. Goyal, R. Girshick, K. He, and P. Doll ´ar, “Focal loss for dense object detection,” in Proceedings of the IEEE International Conference on Computer Vision , 2017, pp. 2980–2988

  27. [35]

    Tversky loss function for image segmentation using 3d fully convolutional deep networks,

    S. S. M. Salehi, D. Erdogmus, and A. Gholipour, “Tversky loss function for image segmentation using 3d fully convolutional deep networks,” in International Workshop on Machine Learning in Medical Imaging . Springer, 2017, pp. 379–387

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.