Pith. sign in

REVIEW 5 major objections 5 minor 130 references

Image Segmentation with Large Language Models: A Survey with Perspectives for Intelligent Transportation Systems

T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This survey argues that LLM-guided image segmentation is reshaping intelligent transportation, but pays a measurable 'cost of generalization' against specialized models.

desk verdict A map of the field worth having, but unsourced benchmark tables and placeholder references sink the central 'cost of generalization' claim as an evidence-based finding. read the letter →

arxiv 2506.14096 v2 pith:D652ZMKG submitted 2025-06-17 cs.CV cs.AI

Sanjeda Akter , Ibne Farabi Shihab , Anuj Sharma This is my paper · ORCID
classification cs.CVcs.AI
keywords LargeLanguageModelsImageSegmentationIntelligentTransportationSystemsVision-LanguageOpen-VocabularyAutonomousDrivingSegmentAnythingModelCostofGeneralization
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 survey argues that combining large language models with image segmentation, a paradigm it calls vision-language segmentation (VLSeg), is moving intelligent transportation systems from fixed class labels toward free-form, instruction-guided scene understanding. Its load-bearing claim is that this flexibility comes at a measurable cost: open-vocabulary VLSeg models trail highly optimized, single-task segmentation models on safety-critical driving benchmarks. The survey organizes the field with a taxonomy based on prompting interfaces and core architectures, and uses comparative tables to quantify both clean performance and degradation under weather, night, blur, and adversarial text prompts. If the claim is right, practitioners should treat language-guided segmentation as a complement to, not a replacement for, specialized perception modules in production vehicles.

What carries the argument

The load-bearing machinery is the pairing of two contrasts: the survey's taxonomy, which sorts VLSeg models by prompting interface and core architecture, and the documented performance gap between open-vocabulary models and a specialized supervised reference, OneFormer, which the paper calls the 'cost of generalization.' Cross-attention fusion, written as $\mathrm{softmax}\left(\frac{Q K^T}{\sqrt{d}}\right) V'$, is presented as the mechanism by which language guides visual feature selection, and as the main computational bottleneck for real-time ITS deployment. Robustness tables add a second axis, quantifying performance degradation under adverse weather, night, motion blur, and adversarial text prompts, with the largest drops in snow and adversarial text.

What would settle it

Run the paper's proposed standardized ITS protocol, with a fixed prompt vocabulary, Cityscapes, BDD100K, and nuScenes subsets, and public corruption sets, on LISA, SEEM, Grounded-SAM, and a specialized supervised model; if the roughly 20-point mIoU gap shrinks to noise or a specialized model no longer tops every dataset, the cost-of-generalization claim loses its evidence base.

Watch

Extended reading notes

Core claim

The paper's central contention is that LLM-augmented vision-language segmentation is transforming perception for autonomous driving, traffic monitoring, and infrastructure maintenance, but that this transformation carries a documented 'cost of generalization.' On the benchmarks it compiles, open-vocabulary models such as LISA reach 47.3 mIoU on Cityscapes while a specialized supervised model, OneFormer, reaches 68.0, a gap of roughly 20 points that persists across KITTI, BDD100K, nuScenes, Waymo Open, and Argoverse 2. The survey defends this claim by building a taxonomy organized by prompting interface (text, geometric, multimodal) and core architecture (vision-language pre-training, promptable foundation models, hybrid detector-segmentation, unified architectures), and by reporting robustness results showing steep drops under snow, night, motion blur, and adversarial text prompts. The practical conclusion is that language-guided models are not a drop-in replacement for closed-set specialists, but an added layer for open-world flexibility, human-AI interaction, and explainable reasoning.

Load-bearing premise

The load-bearing premise is that the numbers in the survey's comparison tables are accurate and comparable across papers, including the robustness table that is presented without its own protocol, code, or data.

Editorial extensions

If this is right

  • Language-guided segmentation should be expected to complement, not replace, specialized perception models for safety-critical classes such as traffic signs and lane markings.
  • Real-time deployment will depend on lightweight variants and efficient attention mechanisms, since full VLSeg models currently exceed the latency budgets of automotive-grade hardware.
  • Robustness under adverse weather and adversarial text prompts is a first-order concern, and current models show severe degradation in snow, night, and misleading language inputs.
  • Evaluation of VLSeg for transportation needs a standardized protocol with multiple ITS datasets, compositional prompts, and metrics beyond mIoU, including boundary quality, small-object IoU, temporal consistency, and inference latency.
  • Modular hybrid systems may be easier to certify under component-level safety standards than end-to-end unified models, which would require new validation methodologies.

Reading between the lines

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

  • If the performance gap persists, production architectures will likely become hybrid: closed-set segmentation for regulated, safety-critical classes, with LLM-guided segmentation reserved for novelty detection, explanation, and human-in-the-loop teleoperation.
  • The paper's proposed standardized evaluation protocol is directly testable: running it on LISA, SEEM, Grounded-SAM, and a specialized supervised baseline would quickly reveal whether the documented ~20-point gap is a genuine property of open-vocabulary generalization or an artifact of dataset-specific fine-tuning.
  • The reported vulnerability to adversarial text prompts suggests the language interface itself is a safety-critical attack surface, implying that input sanitization and prompt validation deserve formal treatment in future designs.
  • For driving safety, temporal consistency across video frames may matter more than single-image mIoU, so extending VLSeg evaluation toward long-horizon tracking benchmarks is a natural next step.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The manuscript is a survey of LLM-augmented image segmentation (VLSeg) with an explicit focus on intelligent transportation systems (ITS). It proposes a taxonomy organized by prompting interface and core architecture, reviews vision and language encoders and fusion strategies, presents comparative performance tables, discusses ITS applications, datasets, challenges, and future directions, and concludes with a call for explainable, human-centric AI. The central thesis is a "cost of generalization": open-vocabulary, language-guided segmentation models offer flexibility but currently lag highly optimized, specialized models on safety-critical ITS tasks, so they should complement rather than replace specialized perception modules.

Significance. If its empirical claims were properly sourced, this survey would be a useful entry point for ITS practitioners and researchers. The taxonomic organization, the mathematical formulation of cross-attention fusion, and the proposed standardized evaluation protocol are constructive contributions. However, the quantitative backbone of the paper, especially Tables IV and V, is not verifiable, and the reference list contains multiple placeholders and malformed entries. Because the "cost of generalization" conclusion rests on these unsupported comparisons, the manuscript in its current form does not provide a dependable systematic review.

major comments (5)
  1. [Section V.A, Table V] Table V reports exact mIoU values and percentage degradations for seven models under rain, fog, night, snow, motion blur, and adversarial text, but the text gives no experimental protocol. There are no dataset splits, checkpoint versions, prompt templates, corruption severity levels, or code release. Without this information, the robustness numbers cannot be reproduced or traced to a source, so the conclusion that specialized models such as ClearVision are more resilient is unsupported.
  2. [Section V.A, Table IV] Table IV lists mIoU for eight models across six ITS datasets (Cityscapes, KITTI, BDD100K, nuScenes, Waymo Open, Argoverse 2) without per-cell citations or any description of how these numbers were obtained. The claim of a consistent twenty-point gap between LISA and OneFormer, which underpins the "cost of generalization" argument, is therefore not verifiable. Including models that are not normally evaluated on these benchmarks (e.g., "SAM + CLIP") makes the table especially difficult to trust.
  3. [References [26], [28], [94], [101], [123], [128]] Several references are placeholders or contain malformed identifiers, including [28] which reads "arXiv=2405.12345", [26] which cites a fabricated IEEE URL, [94] and [101] which are explicitly labeled "placeholder", and [123] which has an invalid arXiv identifier. These citations are used in load-bearing positions, such as the Road-Seg-VL dataset in Sections II and X and the fairness discussion in Section XI. A survey cannot support its comparative claims on unverifiable citations.
  4. [Section V.A and Section III.A] The survey relies on the authors' own preprints as evidence for performance and robustness, including ClearVision [78] in Table V, HybridMamba [79] for temporal localization in traffic footage, and the sidewalk-detection result [77] used to support the "cost of generalization" claim. These are unreviewed preprints without independent reproduction. Treating them as benchmark evidence is circular and does not meet the standard required for a survey's comparative conclusions.
  5. [Section III.A, Table I] The encoder comparison in Table I reports parameters, FLOPs, edge-GPU inference time, memory usage, Cityscapes mIoU, and adverse-condition performance degradation for eight encoders, but no source or evaluation protocol is given. The surrounding claims about real-time suitability and robustness to adverse conditions rely on these unverified numbers.
minor comments (5)
  1. [Section II.D] The phrase "We provide ataxonomy" should read "We provide a taxonomy".
  2. [Section X.B] LISA is described as a "large-scale dataset," but reference [76] is a reasoning-segmentation model paper; the dataset description should be corrected or supported by a separate citation.
  3. [Table II and Section V] The model name "LLaV A-1.5" contains an erroneous space and should be written as "LLaVA-1.5".
  4. [Reference [20]] The title of reference [20] refers to "autonomous driving," but the cited Grounding DINO paper is about open-set object detection; the title should be corrected.
  5. [Figure 3] The label "V oyager" contains an unintended space and should read "Voyager".

Circularity Check

2 steps flagged · score 5.0 of 10

The 'cost of generalization' conclusion leans on same-author preprints for its safety-critical and robustness evidence, while the survey's taxonomy and architecture review remain independent.

  1. self citation load bearing [Section V-A (Quantitative Performance Benchmarking), sidewalk claim following Table III; Refs [77]]
    "This is particularly true for safety-critical sub-tasks like robust sidewalk detection, where specialized ensemble models have been shown to surpass the capabilities of more general LLM-based approaches [77]."

    Ref [77] is a same-author preprint (Shihab, Bhagat, and Sharma; Shihab and Sharma are authors of this survey). The sentence is the paper's only support for the sub-claim that specialized models beat LLM-based approaches on a safety-critical ITS task, and that sub-claim is one of the two pillars of the 'cost of generalization' conclusion. The survey provides no independent benchmark or reproduction for [77]; the claim is therefore imported from the authors' own unverified work rather than from an external result.

  2. self citation load bearing [Section V-A.3, Table V (Empirical Robustness Testing); Refs [78], [79]]
    "Purpose-built models like ClearVision [78], which uses CycleGAN for domain adaptation and SigLIP-2 for robust feature extraction, show significantly better resilience to adverse weather conditions."

    Table V is presented as 'empirical robustness testing' but gives no protocol, code, or per-cell source; its specialized-model row is ClearVision [78], a same-author preprint, and the surrounding text also cites the same authors' HybridMamba [79] for robustness improvements. The survey then uses this table to conclude that specialized models are more robust and that general VLSeg models pay a robustness cost, so the quantitative robustness evidence loops back to the authors' own unreviewed preprints.

full rationale

The paper is a survey rather than a first-principles derivation, so the most common circularity patterns (self-definitional, fitted-input-as-prediction, uniqueness-imported-from-authors, ansatz-smuggled) do not directly apply. The taxonomy in Section IV, the architectural deep dive in Section III, and the cross-attention formulation are self-contained and non-circular. The central 'cost of generalization' claim is a synthesis of benchmark comparisons and does not reduce to a single equation by construction. However, two of its load-bearing evidence pillars are same-author preprints: the sidewalk-specific claim ([77]) and the robustness table's specialized-model comparison ([78], with [79] cited in the text). Because Table V has no protocol and Table IV has no per-cell provenance, the quantitative support for the central claim cannot be independently checked. I am not treating the placeholder or fabricated reference entries (e.g., [26], [28], [94], [101], [123], [128]) as circularity; those are evidence-integrity and correctness risks, not identity-of-input/output reductions. The OneFormer-versus-OpenSeg comparison in Table III also mixes PQ and mIoU, but that is a comparability flaw rather than a circular step. The score of 5 reflects that the survey's comparative conclusion is partially anchored in the authors' own unverified work while retaining independent taxonomic and architectural content.

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

The survey introduces no fitted parameters or invented physical entities, so those ledgers are empty. The load-bearing assumptions are about citation integrity, comparability of benchmark numbers, and the independence of self-cited evidence; all three are questionable given the placeholders and unsourced tables.

assumptions (3)
  • domain assumption The cited references exist, are accurately described, and support the claims attributed to them.
    The survey's synthesis rests on citation integrity; many entries are placeholders or synthetic (e.g., Refs [26], [28], [59], [123]), so this axiom is violated.
  • domain assumption Quantitative benchmark numbers in Tables I, III, IV, and V were measured or reported under protocols comparable enough for the stated comparisons.
    No methodology, code, or data release is provided; Table V presents 'empirical robustness testing' without a protocol, so the comparison is not independently checkable.
  • ad hoc to paper The authors' self-cited preprints provide independent support for the robustness and performance claims.
    ClearVision [78] and HybridMamba [79] are used as benchmarks in Table V and Section V; they are same-author preprints without shipped code or external reproduction, so treating them as external evidence is circular.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Image Segmentation with Large Language Models: A Survey with Perspectives for Intelligent Transportation Systems." pith.science (2026). https://pith.science/paper/D652ZMKG

@misc{pith2026250614096,
  author       = {Pith},
  title        = {Pith review of: Image Segmentation with Large Language Models: A Survey with Perspectives for Intelligent Transportation Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D652ZMKG}},
  note         = {Machine review of arXiv:2506.14096}
}
read the original abstract

The integration of Large Language Models (LLMs) with computer vision is profoundly transforming perception tasks like image segmentation. For intelligent transportation systems (ITS), where accurate scene understanding is critical for safety and efficiency, this new paradigm offers unprecedented capabilities. This survey systematically reviews the emerging field of LLM-augmented image segmentation, focusing on its applications, challenges, and future directions within ITS. We provide a taxonomy of current approaches based on their prompting mechanisms and core architectures, and we highlight how these innovations can enhance road scene understanding for autonomous driving, traffic monitoring, and infrastructure maintenance. Finally, we identify key challenges, including real-time performance and safety-critical reliability, and outline a perspective centered on explainable, human-centric AI as a prerequisite for the successful deployment of this technology in next-generation transportation systems.

Figures

Figures reproduced from arXiv: 2506.14096 by the authors.

Figure 1
Figure 1. Taxonomy of Image Segmentation with Large Language Models for Intelligent Transportation Systems specialized models. Analyzing this trade-off is essential for their practical deployment in ITS. Despite these advancements, integrating LLMs into segmen￾tation for ITS faces several challenges. Real-time performance is a critical bottleneck, as large models like SAM [31] incur high computational costs, necessitating lig… view at source ↗
Figure 2
Figure 2. Timeline of key milestones in the evolution of segmentation for ITS, with a color-coded legend at the top indicating categories. like SegNet [44] and U-Net [45] introduced the powerful encoder-decoder paradigm. The encoder, typically a pre￾trained classification network (e.g., VGG [54]), progressively downsamples the input to capture semantic context, while the decoder upsamples these features to reconstruct a full￾… view at source ↗
Figure 3
Figure 3. A hierarchical categorization of the reviewed literature. Research flows from low-level foundational concepts to mid-level task-specific models, and finally to high-level integrated systems for ITS applications. CLIPSeg [29], OpenSeg [30], and video-based systems like XMem [61], highlighting their impact on ITS ap￾plications such as pedestrian segmentation, traffic sign recognition, and dynamic obstacle avoidance [2… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: A high-level diagram of a generic Vision-Language Segmentation (VLSeg) pipeline. An image and a text prompt are processed by their respective encoders, fused in a multi-modal module, and then passed to a decoder to generate the final mask. TABLE I: Comparative Analysis…
Figure 5
Figure 5. Figure 5: Computational efficiency comparison of VLSeg models. The chart shows inference time and model size for different models, highlighting the trade-off between performance and computational requirements. IV. A TAXONOMY OF LLM-AUGMENTED SEGMENTATION FOR ITS The rapid integr…
Figure 6
Figure 6. Figure 6: Comparison of Multi-Modal Fusion Mechanisms in VLSeg for ITS Applications. Each strategy presents a different trade-off between computational cost, semantic richness, and suitability for real-time systems. operation proceeds as follows: First, the text and image featur…
Figure 7
Figure 7. Figure 7: Application scenarios of Vision-Language Segmentation (VLSeg) in Intelligent Transportation Systems (ITS). Each panel illustrates how a natural language prompt can be used to segment specific objects of interest in a real-world scene, highlighting the practical utility…
Figure 8
Figure 8. Figure 8: Conceptual diagram of an end-to-end reasoning loop in ITS. The VLSeg module provides scene understanding to a central LLM, which processes goals and generates both actionable commands for the vehicle and human-understandable explanations. by LLMs, presents a highly eff…
Figure 9
Figure 9. Figure 9: Analysis of common failure modes for VLSeg in ITS. Each panel shows a vulnerability, its safety impact, and a mitigation. To systematically evaluate and compare the performance of segmentation models, a standardized set of metrics is employed. • Intersection-over-Union…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

130 extracted references · 31 canonical work pages

  1. [28]

    Road-seg-vl: Vision-language dataset for road scene segmentation,

    N. specified, “Road-seg-vl: Vision-language dataset for road scene segmentation,”arXiv preprint arXiv=2405.12345, 2024

    arXiv 2024

  2. [94]

    A review of deep learning- based methods for pavement defect detection,

    K. Zhang, Y . Zhang, and H. D. Cheng, “A review of deep learning- based methods for pavement defect detection,”IEEE Transactions on Intelligent Transportation Systems, 2024, placeholder, represents the body of literature on this topic

    work page 2024

  3. [101]

    Lingo-1: A foundation model for language-driven autonomous vehicles,

    W. Technologies, “Lingo-1: A foundation model for language-driven autonomous vehicles,”Wayve Blog, 2024. [Online]. Available: https://wayve.com/blog/ lingo-1-foundation-model-language-driven-autonomous-vehicles/

    work page 2024

  4. [128]

    Vlm-ad: End-to-end autonomous driving through vision-language model supervision,

    Y . Xu, Y . Hu, Z. Zhang, G. P. Meyer, S. K. Mustikovela, S. Srinivasa, E. M. Wolff, and X. Huang, “Vlm-ad: End-to-end autonomous driving through vision-language model supervision,” 2024. [Online]. Available: https://arxiv.org/abs/2412.14446

    arXiv 2024

  5. [26]

    Traffic scene analysis using vision-language models,

    L. Wanget al., “Traffic scene analysis using vision-language models,”IEEE Transactions on vehicular technology, 2023, based on context; aligned with similar entries like wang2024smartcity. [Online]. Available: https://www.ieee.org/abs/10234567

    work page arXiv 2023

  6. [123]

    A continual learning survey: Defying forgetting in classification and deep learning,

    M. De Lange, R. Aljundi, M. Masana, S. Parisot, X. Jia, A. Leonardis, G. Slabaugh, and T. Tuytelaars, “A continual learning survey: Defying forgetting in classification and deep learning,”IEEE transactions on pattern analysis and machine intelligence, vol. 44, no. 7, pp. 3366– 3385, 2021

    work page 2021

  7. [78]

    Bilinear attention net- works,

    J.-H. Kim, J. Kim, H. Cha, and J.-W. Lee, “Bilinear attention net- works,” inAdvances in Neural Information Processing Systems, 2018

    2018

  8. [79]

    Crash time matters: Hybridmamba for fine-grained temporal localization in traffic surveillance footage,

    I. F. Shihab and A. Sharma, “Crash time matters: Hybridmamba for fine-grained temporal localization in traffic surveillance footage,”arXiv preprint arXiv:2504.03235, 2025. IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. XX, NO. X, MONTH YEAR 24

    arXiv 2025

  9. [77]

    Precise and Robust Sidewalk Detection: Leveraging Ensemble Learning to Surpass LLM Limitations in Urban Environments

    I. F. Shihab, S. R. Bhagat, and A. Sharma, “Precise and robust sidewalk detection: Leveraging ensemble learning to surpass llm limitations in urban environments,”arXiv preprint arXiv:2405.14876, 2024

    work page Pith review arXiv 2024

Show all 130 references
  1. [1]

    Vision language models in autonomous driving: A survey and outlook,

    X. Zhou, M. Liu, E. Yurtsever, B. L. Zagar, W. Zimmer, H. Cao, and A. C. Knoll, “Vision language models in autonomous driving: A survey and outlook,”arXiv preprint arXiv:2310.14414, 2024. [Online]. Available: https://arxiv.org/abs/2310.14414

    2024 arXiv

  2. [2]

    Multi-modal sensor fusion for auto driving perception: A survey,

    K. Huang, B. Shi, X. Li, X. Li, S. Huang, and Y . Li, “Multi-modal sensor fusion for auto driving perception: A survey,” arXiv preprint arXiv:2202.02703, 2024. [Online]. Available: https: //arxiv.org/abs/2202.02703

    2024 arXiv

  3. [3]

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

    L.-C. Chenet al., “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

    2017

  4. [4]

    Mask r-cnn,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask r-cnn,”IEEE International Conference on Computer Vision (ICCV), pp. 2961–2969, 2017

    2017

  5. [5]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liuet al., “Swin transformer: Hierarchical vision transformer using shifted windows,”IEEE International Conference on Computer Vision (ICCV), pp. 10 012–10 022, 2021

    2021

  6. [6]

    Segmenter: Transformer for semantic segmentation,

    R. Strudelet al., “Segmenter: Transformer for semantic segmentation,” IEEE International Conference on Computer Vision (ICCV), vol. –, pp. 7262–7272, 2021

    2021

  7. [7]

    The cityscapes dataset for semantic urban scene understanding,

    M. Cordts, M. Omran, S. Ramos, T. Rehfeld, M. Enzweiler, R. Benenson, U. Franke, S. Roth, and B. Schiele, “The cityscapes dataset for semantic urban scene understanding,” 2016. [Online]. Available: https://arxiv.org/abs/1604.01685

    2016 arXiv

  8. [8]

    Bdd100k: A diverse driving dataset for heterogeneous multitask learning,

    F. Yuet al., “Bdd100k: A diverse driving dataset for heterogeneous multitask learning,”IEEE Conference on Computer Vision and Pattern Recognition (CVPR), vol. IEEE, pp. 2636–2645, 2020

    2020

  9. [9]

    Encoder-decoder with atrous separable convolution for semantic image segmentation,

    L.-C. Chenet al., “Encoder-decoder with atrous separable convolution for semantic image segmentation,”European Conference on Computer Vision (ECCV), pp. 801–818, 2018

    2018

  10. [10]

    Panoptic segmenta- tion,

    A. Kirillov, R. Girshick, K. He, and P. Doll ´ar, “Panoptic segmenta- tion,”IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 9404–9413, 2019

    2019

  11. [11]

    Deep learning for 3d point clouds: A survey,

    Y . Guoet al., “Deep learning for 3d point clouds: A survey,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. IEEE, no. 12, pp. 4338–4364, 2020, assumed to be the intended reference for semantic segmentation; actual title may vary

    2020

  12. [12]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiyet al., “An image is worth 16x16 words: Transformers for image recognition at scale,”International Conference on Learning Representations (ICLR), 2021. [Online]. Available: https: //arxiv.org/abs/2010.11929

    2021 arXiv

  13. [13]

    nuscenes: A multimodal dataset for autonomous driv- ing,

    H. Caesaret al., “nuscenes: A multimodal dataset for autonomous driv- ing,”IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 11 621–11 631, 2020

    2020

  14. [14]

    The mapillary vistas dataset for semantic under- standing of street scenes,

    G. Neuholdet al., “The mapillary vistas dataset for semantic under- standing of street scenes,”IEEE International Conference on Computer Vision (ICCV), pp. 4990–4999, 2017

    2017

  15. [15]

    Bert: Pre-training of deep bidirectional transformers for language understanding,

    J. Devlinet al., “Bert: Pre-training of deep bidirectional transformers for language understanding,”arXiv preprint arXiv:1810.04805, 2018

    2018 arXiv

  16. [16]

    Language models are few-shot learners,

    T. Brownet al., “Language models are few-shot learners,”Advances in Neural Information Processing Systems (NeurIPS), vol. 33, pp. 1877–1901, 2020. [Online]. Available: https://arxiv.org/abs/2005.14165

    1901 arXiv

  17. [17]

    Exploring the limits of transfer learning with a unified text-to-text transformer,

    C. Raffelet al., “Exploring the limits of transfer learning with a unified text-to-text transformer,”Journal of Machine Learning Research, vol. 21, pp. 1–67, 2020. [Online]. Available: https: //arxiv.org/abs/1910.10683

    2020 arXiv

  18. [18]

    Learning transferable visual models from natural language supervision,

    A. Radfordet al., “Learning transferable visual models from natural language supervision,”International Conference on Machine Learning (ICML), vol. PMLR, pp. 8748–8763, 2021

    2021

  19. [19]

    Dinov2: Learning robust visual features without supervision,

    M. Oquabet al., “Dinov2: Learning robust visual features without supervision,”arXiv preprint arXiv:2304.07193, 2023. [Online]. Available: https://arxiv.org/abs/2304.07193

    2023 arXiv

  20. [21]

    Segment everything everywhere all at once,

    X. Zouet al., “Segment everything everywhere all at once,” arXiv preprint arXiv:2304.06718, 2023. [Online]. Available: https: //arxiv.org/abs/2304.06718

    2023 arXiv

  21. [23]

    Drivelm: Driving with language and vision models,

    A. Simaet al., “Drivelm: Driving with language and vision models,”arXiv preprint arXiv:2402.12289, 2024. [Online]. Available: https://arxiv.org/abs/2402.12289

    2024 arXiv

  22. [24]

    Traffic scene perception via multimodal large language model with data augmentation and efficient training strategy,

    T. Chenet al., “Traffic scene perception via multimodal large language model with data augmentation and efficient training strategy,”Expert Systems with Applications, 2024. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0957417424018244

    2024

  23. [25]

    Exploring the roles of large language models in reshaping transportation systems: A survey, framework, and roadmap,

    K. Liet al., “Exploring the roles of large language models in reshaping transportation systems: A survey, framework, and roadmap,”arXiv preprint arXiv:2403.17608, 2024. [Online]. Available: https://arxiv.org/abs/2403.17608

    2024 arXiv

  24. [27]

    Talk2car: Taking control of your self-driving car,

    T. Deruytter and S. Vandenhende, “Talk2car: Taking control of your self-driving car,”European Conference on Computer Vision (ECCV), vol. Springer, no. 091-, pp. 797–814, 2019

    2019

  25. [29]

    Image segmentation using text and image prompts,

    T. L ¨uddecke and A. Ecker, “Image segmentation using text and image prompts,”IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7086–7096, 2022

    2022

  26. [30]

    Scaling open-vocabulary image segmentation with image-level labels,

    G. Ghiasiet al., “Scaling open-vocabulary image segmentation with image-level labels,”European Conference on Computer Vision (ECCV), vol. 136, pp. 540–557, 2022

    2022

  27. [31]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo, P. Doll ´ar, and R. Girshick, “Segment anything,” 2023. [Online]. Available: https://arxiv.org/abs/2304.02643

    2023 arXiv

  28. [32]

    Faster segment anything: Towards lightweight sam for mobile applications,

    C. Zhang, D. Han, Y . Qiao, J. U. Kim, S.-H. Bae, S. Lee, and C. S. Hong, “Faster segment anything: Towards lightweight sam for mobile applications,”arXiv preprint arXiv:2306.14289, 2023

    2023 arXiv

  29. [33]

    Edgevits: Competing light-weight cnns on mobile devices with vision transformers,

    J. Pan, A. Bulat, F. Tan, X. Zhu, L. Dudziak, H. Li, G. Tzimiropoulos, and B. Martinez, “Edgevits: Competing light-weight cnns on mobile devices with vision transformers,” 2022. [Online]. Available: https://arxiv.org/abs/2205.03436

    2022 arXiv

  30. [34]

    Robust image classification with multi- modal large language models,

    H. Wanget al., “Robust image classification with multi- modal large language models,”Pattern Recognition Letters, 2023. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0167865525001618 IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. XX, N...

    2023

  31. [36]

    Driving forward: Semantic segmenta- tion in autonomous vehicles,

    Keylabs, “Driving forward: Semantic segmenta- tion in autonomous vehicles,”Keylabs AI Blog,

  32. [37]

    Real-time semantic segmentation for autonomous driving: A review of cnns, transformers, and beyond,

    M. A. Elhassan, C. Zhou, A. Khan, A. Benabid, A. B. Adam, A. Mehmood, and N. Wambugu, “Real-time semantic segmentation for autonomous driving: A review of cnns, transformers, and beyond,” Journal of King Saud University-Computer and Information Sciences, p. 102226, 2024

    2024

  33. [38]

    Available: https://arxiv.org/abs/2405.11234

    [Online]. Available: https://arxiv.org/abs/2405.11234

    arXiv

  34. [39]

    Real-time vehicle detection and tracking for autonomous driving,

    N. specified, “Real-time vehicle detection and tracking for autonomous driving,”Journal of Autonomous Vehicles, 2023. [Online]. Available: https://arxiv.org/abs/2305.13457

    2023 arXiv

  35. [40]

    Available: https://www.keylabs.ai/blog/ driving-forward-semantic-segmentation-in-autonomous-vehicles/

    [Online]. Available: https://www.keylabs.ai/blog/ driving-forward-semantic-segmentation-in-autonomous-vehicles/

  36. [41]

    Oneformer: One transformer to rule them all for universal image segmentation,

    J. Jainet al., “Oneformer: One transformer to rule them all for universal image segmentation,”IEEE Conference on Computer Vision and Pattern Recognition (CVPR), vol. 1-6656-6643, pp. 16 471–16 481, 2022

    2022

  37. [42]

    Faster r-cnn: Towards real-time object detection with region proposal networks,

    S. Renet al., “Faster r-cnn: Towards real-time object detection with region proposal networks,”Advances in Neural Information Processing Systems (NeurIPS), vol. 28, pp. 91–99, 2015. [Online]. Available: https://arxiv.org/abs/1506.01497

    2015 arXiv

  38. [43]

    Fully convolutional networks for semantic segmentation,

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

    2015

  39. [44]

    Masked-attention mask transformer for universal im- age segmentation,

    B. Chenget al., “Masked-attention mask transformer for universal im- age segmentation,”IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1290–1299, 2022

    2022

  40. [45]

    U-net: Convolutional net- works for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional net- works for biomedical image segmentation,”International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), vol. Springer, pp. 234–241, 2015

    2015

  41. [46]

    Normalized cuts and image segmentation,

    J. Shi and J. Malik, “Normalized cuts and image segmentation,”IEEE Transactions on pattern analysis and machine intelligence, vol. 22, no. 8, pp. 888–905, 2000

    2000

  42. [47]

    Cross-modal self-attention network for referring image segmentation,

    L. Ye, M. Rochan, Z. Liu, and Y . Wang, “Cross-modal self-attention network for referring image segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 10 502–10 511

    2019

  43. [48]

    Segnet: A deep con- volutional encoder-decoder architecture for image segmentation,

    V . Badrinarayanan, A. Kendall, and R. Cipolla, “Segnet: A deep con- volutional encoder-decoder architecture for image segmentation,”IEEE transactions on pattern analysis and machine intelligence, vol. 39, no. 12, pp. 2481–2495, 2017

    2017

  44. [49]

    Phrasecut: Language-based image segmentation in the wild,

    C. Wu, C. Feichtenhofer, H. Fan, K. He, P. Krahenbuhl, and R. Gir- shick, “Phrasecut: Language-based image segmentation in the wild,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 10 216–10 225

    2020

  45. [50]

    Segformer: Simple and efficient design for semantic segmentation with transformers,

    E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “Segformer: Simple and efficient design for semantic segmentation with transformers,” inAdvances in Neural Information Processing Systems, vol. 34, 2021, pp. 12 077–12 090

    2021

  46. [51]

    Vilbert: Pretraining task- agnostic visiolinguistic representations for vision-and-language tasks,

    J. Lu, D. Batra, D. Parikh, and S. Lee, “Vilbert: Pretraining task- agnostic visiolinguistic representations for vision-and-language tasks,” inAdvances in Neural Information Processing Systems, 2019, pp. 13– 23

    2019

  47. [52]

    Bi-directional re- lationship inferring network for referring image segmentation,

    Z. Hu, G. Feng, J. Sun, L. Zhang, and H. Lu, “Bi-directional re- lationship inferring network for referring image segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 4424–4433

    2020

  48. [53]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” inAdvances in neural information processing systems 25, 2012

    2012

  49. [54]

    Refvos: A closer look at referring expressions for video object segmentation,

    M. Bellver, X. Gir ´o-i Nieto, F. Marqu ´es, and J. Torres, “Refvos: A closer look at referring expressions for video object segmentation,” in Proceedings of the 28th ACM International Conference on Multimedia, 2020, pp. 3569–3578

    2020

  50. [55]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, “Attention is all you need,”

  51. [56]

    Lxmert: Learning cross-modality encoder rep- resentations from transformers,

    H. Tan and M. Bansal, “Lxmert: Learning cross-modality encoder rep- resentations from transformers,” inProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, 2019, pp. 5100–5111

    2019

  52. [57]

    Llm-seg: Bridging image segmentation with large language models reasoning,

    J. Yuet al., “Llm-seg: Bridging image segmentation with large language models reasoning,”arXiv preprint arXiv:2404.08767, 2024. [Online]. Available: https://arxiv.org/abs/2404.08767

    2024 arXiv

  53. [58]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,”arXiv preprint arXiv:1409.1556, 2014

    2014 arXiv

  54. [59]

    A survey on vision-language foundation models,

    J. Liet al., “A survey on vision-language foundation models,” arXiv preprint arXiv:2305.12345, 2023. [Online]. Available: https: //arxiv.org/abs/2305.12345

    2023 arXiv

  55. [60]

    Available: https://arxiv.org/abs/1706.03762

    [Online]. Available: https://arxiv.org/abs/1706.03762

    arXiv

  56. [61]

    Semantic segmentation datasets for autonomous driving,

    N. specified, “Semantic segmentation datasets for autonomous driving,”Medium, 2023. [On- line]. Available: https://medium.com/hackernoon/subject/ semantic-segmentation-datasets-for-autonomous-driving-118217ebd2aff0

    2023

  57. [62]

    Open-vocabulary vision-language segmentation for autonomous driving,

    H. Zhanget al., “Open-vocabulary vision-language segmentation for autonomous driving,”arXiv preprint arXiv:2404.17890, 2024. [Online]. Available: https://arxiv.org/abs/2404.17890

    2024 arXiv

  58. [63]

    A survey on multimodal large language models for autonomous driving,

    C. Cui, Y . Ma, X. Cao, W. Ye, Y . Zhou, K. Liang, J. Chen, J. Lu, Z. Yang, K.-D. Liao, T. Gao, E. Li, K. Tang, Z. Cao, T. Zhou, A. Liu, X. Yan, S. Mei, J. Cao, Z. Wang, and C. Zheng, “A survey on multimodal large language models for autonomous driving,”arXiv preprint arXiv:23...

    2023 arXiv

  59. [64]

    Are we ready for autonomous driving? the kitti vision benchmark suite,

    A. Geiger, P. Lenz, and R. Urtasun, “Are we ready for autonomous driving? the kitti vision benchmark suite,”IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), pp. 3354–3361, 2012

    2012

  60. [65]

    Architectures for vision-language segmentation: A survey,

    W. Chenet al., “Architectures for vision-language segmentation: A survey,”arXiv preprint arXiv:2405.09812, 2024. [Online]. Available: https://arxiv.org/abs/2405.09812

    2024 arXiv

  61. [66]

    Xmem: Long-term video object segmentation with an atkinson-shiffrin memory model,

    H. K. Cheng and A. G. Schwing, “Xmem: Long-term video object segmentation with an atkinson-shiffrin memory model,” 2022. [Online]. Available: https://arxiv.org/abs/2207.07115

    2022 arXiv

  62. [67]

    Efficient unstructured pruning of mamba state-space models for resource-constrained environments,

    I. F. Shihab, S. Akter, and A. Sharma, “Efficient unstructured pruning of mamba state-space models for resource-constrained environments,” arXiv preprint arXiv:2505.08299, 2025

    2025

  63. [68]

    Vlm-ad: Vision-language models for autonomous driving,

    L. Zhanget al., “Vlm-ad: Vision-language models for autonomous driving,”arXiv preprint arXiv:2404.14567, 2024. [Online]. Available: https://arxiv.org/abs/2404.14567

    2024 arXiv

  64. [69]

    P-tuning: Prompt tuning can be as good as fine-tuning on language models,

    X. Liu, Y . Zheng, Z. Du, M. Ding, Y . Qian, Z. Yang, and J. Tang, “P-tuning: Prompt tuning can be as good as fine-tuning on language models,”arXiv preprint arXiv:2103.10385, 2021

    2021 arXiv

  65. [70]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

    2016

  66. [71]

    Efficientvit: Memory efficient vision transformer with cascaded group attention,

    X. Liu, H. Peng, N. Zheng, Y . Yang, H. Hu, and Y . Yuan, “Efficientvit: Memory efficient vision transformer with cascaded group attention,” pp. 14 420–14 430, 2023

    2023

  67. [72]

    Multimodal compact bilinear pooling for visual ques- tion answering and visual grounding,

    A. Fukui, D. H. Park, D. Yang, A. Rohrbach, T. Darrell, and M. Rohrbach, “Multimodal compact bilinear pooling for visual ques- tion answering and visual grounding,” inProceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, 2016, pp. 457–468

    2016

  68. [73]

    The power of scale for parameter-efficient prompt tuning,

    B. Lester, R. Al-Rfou, and N. Constant, “The power of scale for parameter-efficient prompt tuning,”arXiv preprint arXiv:2104.08691, 2021

    2021 arXiv

  69. [74]

    Unifying vision-and-language tasks via text genera- tion,

    J.-C. Heet al., “Unifying vision-and-language tasks via text genera- tion,”arXiv preprint arXiv:2102.02779, 2022

    2022 arXiv

  70. [75]

    Unify, align and refine: A unified framework for vision-and-language pre-training,

    J. Hu, D. Li, J. Li, S. Savarese, and S. Hoi, “Unify, align and refine: A unified framework for vision-and-language pre-training,” inAdvances in Neural Information Processing Systems, 2021

    2021

  71. [76]

    Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,

    J. Li, D. Li, S. Savarese, and S. Hoi, “Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,” 2023. [Online]. Available: https://arxiv.org/abs/2301.12597

    2023 arXiv

  72. [80]

    Llava-1.5: Improved baselines for visual instruction tuning,

    H. Liu, C. Li, Y . Li, and Y . J. Lee, “Llava-1.5: Improved baselines for visual instruction tuning,”arXiv preprint arXiv:2310.03744, 2023

    2023 arXiv

  73. [81]

    Lisa: Reasoning segmentation via large language models,

    X. Lai, Z. Tian, Y . Shi, and J. Chen, “Lisa: Reasoning segmentation via large language models,”arXiv preprint arXiv:2308.00692, 2023

    2023 arXiv

  74. [82]

    Lidarclip or: How i learned to talk to point clouds,

    G. Pfreundschuh, Y . Che, Z. Murez, C. Liu, G. Gaurav, S. Kumar, J. J. Park, and D. Ramanan, “Lidarclip or: How i learned to talk to point clouds,”arXiv preprint arXiv:2308.00194, 2023

    2023 arXiv

  75. [83]

    Clearvision: Leveraging cyclegan and siglip-2 for robust all- weather classification in traffic camera imagery,

    A. L. Sivaraman, K. Adu-Gyamfi, I. F. Shihab, and A. Sharma, “Clearvision: Leveraging cyclegan and siglip-2 for robust all- weather classification in traffic camera imagery,”arXiv preprint arXiv:2504.19684, 2025

    2025

  76. [84]

    Lg3d: A language- guided 3d object detection framework for autonomous driving,

    J. Ding, J. Zhang, D. Xu, W. Chen, and L. Zhang, “Lg3d: A language- guided 3d object detection framework for autonomous driving,”arXiv preprint arXiv:2212.08411, 2022

    2022 arXiv

  77. [85]

    Pointnet: Deep learning on point sets for 3d classification and segmentation,

    C. R. Qi, H. Su, K. Mo, and L. J. Guibas, “Pointnet: Deep learning on point sets for 3d classification and segmentation,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 652–660

    2017

  78. [86]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space,

    C. R. Qi, L. Yi, H. Su, and L. J. Guibas, “Pointnet++: Deep hierarchical feature learning on point sets in a metric space,” inAdvances in neural information processing systems, vol. 30, 2017

    2017

  79. [87]

    Communication-efficient learning of deep networks from decentral- ized data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentral- ized data,” inArtificial intelligence and statistics. PMLR, 2017, pp. 1273–1282

    2017

  80. [88]

    Clip2scene: Towards label-efficient 3d scene understanding by clip,

    R. You, X. Yan, K. He, P. Li, X. Liu, Y . Liao, Z. Liao, and H. Li, “Clip2scene: Towards label-efficient 3d scene understanding by clip,” inInternational Conference on Learning Representations (ICLR), 2023

    2023

  81. [89]

    Federated learning: Challenges, methods, and future directions,

    T. Li, A. K. Sahu, A. Talwalkar, and V . Smith, “Federated learning: Challenges, methods, and future directions,”IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50–60, 2020

    2020

  82. [90]

    Univs: Unified and universal video segmentation with prompts as queries,

    M. Li, S. Li, X. Zhang, and L. Zhang, “Univs: Unified and universal video segmentation with prompts as queries,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024. [Online]. Available: https://arxiv.org/abs/2402.18115

    2024 arXiv

  83. [91]

    Video object segmentation: A survey,

    L.-J. Yanet al., “Video object segmentation: A survey,”ACM Comput- ing Surveys (CSUR), vol. 54, no. 9, pp. 1–37, 2022

    2022

  84. [92]

    Towards explainable traffic flow prediction with large language models,

    X. Guo, Q. Zhang, J. Jiang, M. Peng, M. Zhu, Hao, and Yang, “Towards explainable traffic flow prediction with large language models,” 2024. [Online]. Available: https://arxiv.org/abs/2404.02937

    2024

  85. [93]

    Advances and open problems in federated learning,

    P. Kairouz, H. B. McMahan, B. Avent, A. Bellet, M. Bennis, A. N. Bhagoji, K. Bonawitz, Z. Charles, G. Cormode, R. Cummingset al., “Advances and open problems in federated learning,”Foundations and Trends® in Machine Learning, vol. 14, no. 1–2, pp. 1–210, 2021

    2021

  86. [95]

    Cooper: Cooperative perception for autonomous driving,

    R. Chen, Z. Li, Y . Hu, Q. Xu, J. Li, and J. Liu, “Cooper: Cooperative perception for autonomous driving,”arXiv preprint arXiv:1904.09333, 2019

    1904 arXiv

  87. [96]

    Cooper: A query-based collaborative perception framework for 3d object detection,

    W. Qin, Z. Wang, L. Chen, J. Li, and J. Liu, “Cooper: A query-based collaborative perception framework for 3d object detection,” in2021 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2021, pp. 12 824–12 830

    2021

  88. [97]

    Driveadapter: A language-grounded driving system via cross-modal feature-conditioned adaptation,

    Y . Fu, Z. Zhang, H. Li, R. Yu, J. Wang, and Z. Sun, “Driveadapter: A language-grounded driving system via cross-modal feature-conditioned adaptation,”arXiv preprint arXiv:2403.09935, 2024

    2024 arXiv

  89. [98]

    Robust and precise sidewalk detection with ensemble learning: Enhancing road safety and facilitating curb space management,

    I. F. Shihab, S. R. Bhagat, and A. Sharma, “Robust and precise sidewalk detection with ensemble learning: Enhancing road safety and facilitating curb space management,” in2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC). IEEE, 2023, pp. 5092–5099

    2023

  90. [99]

    Reason2drive: To- wards interpretable and safe autonomous driving,

    L. Wen, S. Wu, Y . Liu, and J. Luo, “Reason2drive: To- wards interpretable and safe autonomous driving,”arXiv preprint arXiv:2310.08226, 2023

    2023 arXiv

  91. [100]

    Road pothole detection and classification using deep convolutional neu- ral networks,

    H. Maeda, Y . Sekimoto, T. Seto, T. Kashiyama, and H. Omata, “Road pothole detection and classification using deep convolutional neu- ral networks,”Computer-Aided Civil and Infrastructure Engineering, vol. 33, no. 12, pp. 1027–1041, 2018

    2018

  92. [102]

    Talk2bev: Language- grounded bird’s-eye-view for autonomous driving,

    Y . Li, K. Li, L. Pan, and J. Zhang, “Talk2bev: Language- grounded bird’s-eye-view for autonomous driving,”arXiv preprint arXiv:2403.11545, 2024

    2024

  93. [103]

    Lmdrive: Closed-loop end-to-end driving with large language models,

    D. Kimet al., “Lmdrive: Closed-loop end-to-end driving with large language models,”arXiv preprint arXiv:2312.07488, 2023

    2023 arXiv

  94. [104]

    Autort: Neural architecture search for real- time autonomous driving,

    J. Wanget al., “Autort: Neural architecture search for real- time autonomous driving,”arXiv preprint arXiv:2405.09876, 2024. [Online]. Available: https://arxiv.org/abs/2405.09876

    2024 arXiv

  95. [105]

    Driving with language: A unified approach for interpretable and controllable autonomous driving,

    Y . Chenet al., “Driving with language: A unified approach for interpretable and controllable autonomous driving,”arXiv preprint arXiv:2310.05304, 2023

    2023 arXiv

  96. [106]

    Tackling long-tailed instance segmentation with one- shot learning,

    A. Chenet al., “Tackling long-tailed instance segmentation with one- shot learning,”arXiv preprint arXiv:2305.01234, 2023, placeholder representing research on this topic

    2023 arXiv

  97. [107]

    A survey of quantization methods for efficient neural network infer- ence,

    A. Gholami, S. Kim, Z. Dong, Z. Yao, M. W. Mahoney, and K. Keutzer, “A survey of quantization methods for efficient neural network infer- ence,”arXiv preprint arXiv:2103.13630, 2021

    2021 arXiv

  98. [108]

    V oyager: An open-ended embodied agent with large language models,

    G. Wang, Y . Xie, Y . Jiang, A. Mandlekar, C. Xiao, Y . Zhu, L. Fan, and A. Anandkumar, “V oyager: An open-ended embodied agent with large language models,”arXiv preprint arXiv:2305.16291, 2023

    2023 arXiv

  99. [110]

    Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding,

    S. Han, H. Mao, and W. J. Dally, “Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding,”arXiv preprint arXiv:1510.00149, 2015

    2015 arXiv

  100. [111]

    Distilling the knowledge in a neural network,

    G. Hinton, O. Vinyals, and J. Dean, “Distilling the knowledge in a neural network,”arXiv preprint arXiv:1503.02531, 2015

    2015 arXiv

  101. [112]

    Gaia: a generative ai for automotive and industrial applications,

    E. Huet al., “Gaia: a generative ai for automotive and industrial applications,”arXiv preprint arXiv:2310.19685, 2023

    2023 arXiv

  102. [113]

    A comprehensive survey on hardware- aware neural architecture search,

    H. Benmeziane, K. E. Maghraoui, H. Ouarnoughi, S. Niar, M. Wistuba, and N. Wang, “A comprehensive survey on hardware- aware neural architecture search,” 2021. [Online]. Available: https: //arxiv.org/abs/2101.09336

    2021 arXiv

  103. [114]

    Scene parsing through ade20k dataset,

    B. Zhouet al., “Scene parsing through ade20k dataset,”IEEE Confer- ence on Computer Vision and Pattern Recognition (CVPR), no. 1, pp. 633–641, 2017

    2017

  104. [115]

    Available: https://arxiv.org/abs/2405.16789

    [Online]. Available: https://arxiv.org/abs/2405.16789

    arXiv

  105. [116]

    Smart city traffic management with vision-language segmentation,

    L. Wanget al., “Smart city traffic management with vision-language segmentation,”IEEE Transactions on Smart Cities, 2024. [Online]. Available: https://www.ieee.org/abs/105678901

    2024

  106. [117]

    Synthcity: A large scale synthetic point cloud,

    D. Griffiths and J. Boehm, “Synthcity: A large scale synthetic point cloud,” 2019. [Online]. Available: https://arxiv.org/abs/1907.04758

    2019 arXiv

  107. [118]

    Using multimodal large language models for automated detection of traffic safety critical events,

    M. A. Tami, H. I. Ashqar, and M. Elhenawy, “Using multimodal large language models for automated detection of traffic safety critical events,” 2024. [Online]. Available: https://arxiv.org/abs/2406.13894

    2024 arXiv

  108. [119]

    Microsoft coco: Common objects in context,

    T.-Y . Linet al., “Microsoft coco: Common objects in context,”Eu- ropean Conference on Computer Vision (ECCV), vol. Springer, pp. 740–755, 2014

    2014

  109. [120]

    Towards deep learning models resistant to adversarial attacks,

    A. Madry, A. Makelov, L. Schmidt, D. Tsipras, and A. Vladu, “Towards deep learning models resistant to adversarial attacks,”arXiv preprint arXiv:1706.06083, 2017

    2017 arXiv

  110. [121]

    Active learning literature survey,

    B. Settles, “Active learning literature survey,” University of Wisconsin- Madison, Tech. Rep. 1648, 2009

    2009

  111. [122]

    Continualvls: Continual learning for vision-language segmentation in its,

    C. Wanget al., “Continualvls: Continual learning for vision-language segmentation in its,”arXiv preprint arXiv:2405.16792, 2024

    2024 arXiv

  112. [124]

    A survey of methods for explaining black box models,

    R. Guidotti, A. Monreale, S. Ruggieri, F. Turini, F. Giannotti, and D. Pedreschi, “A survey of methods for explaining black box models,” ACM computing surveys (CSUR), vol. 51, no. 5, pp. 1–42, 2018

    2018

  113. [125]

    Reluplex: An efficient smt solver for verifying deep neural networks,

    G. Katz, C. Barrett, D. L. Dill, K. Julian, and M. J. Kochenderfer, “Reluplex: An efficient smt solver for verifying deep neural networks,” inInternational Conference on Computer Aided Verification. Springer, 2017, pp. 97–117

    2017

  114. [126]

    Vision-language models for traffic monitoring and incident detection,

    W. Chenet al., “Vision-language models for traffic monitoring and incident detection,”IEEE Transactions on Intelligent Transportation Systems, 2024. [Online]. Available: https://www.ieeeexplore.ieee.org/ document/10498766

    2024

  115. [127]

    Safeseg: Safe segmentation for high-risk scenarios in autonomous driving,

    W. Liet al., “Safeseg: Safe segmentation for high-risk scenarios in autonomous driving,”arXiv preprint arXiv:2405.13459, 2024. [Online]. Available: https://arxiv.org/abs/2405.13459

    2024 arXiv

  116. [129]

    Explainable ai for autonomous driving: A survey,

    J. Yanget al., “Explainable ai for autonomous driving: A survey,”arXiv preprint arXiv:2405.1000, arXiv preprint arXiv:2405.12347, arXiv

  117. [131]

    Multi-sensor fusion for vision-language segmentation in its,

    C. Wanget al., “Multi-sensor fusion for vision-language segmentation in its,”arXiv preprint arXiv:2405.16790, 2024

    2024 arXiv

  118. [133]

    Unbiased look at dataset bias,

    A. Torralba and A. A. Efros, “Unbiased look at dataset bias,”CVPR 2011, pp. 1521–1528, 2011

    2011

  119. [134]

    A survey on fairness in computer vision,

    J. Zhang, W. Hua, and S. Liu, “A survey on fairness in computer vision,”arXiv preprint arXiv:2401.14 fairness, 2024, placeholder, represents the body of literature on this topic

    2024

  120. [2023]

    Available: https://arxiv.org/abs/2303.05499

    [Online]. Available: https://arxiv.org/abs/2303.05499

    arXiv

  121. [2024]

    Available: https://arxiv.org/abs/2412.14088

    [Online]. Available: https://arxiv.org/abs/2412.14088

    arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.