pith. sign in

arxiv: 2606.08029 · v1 · pith:5MT7SEWQnew · submitted 2026-06-06 · 💻 cs.RO

IntentNav: Learning Spatial-Visual Object Navigation from Human Demonstrations

Yuxin Cai , Zongtai Li , Maonan Wang , Muyi Bao , Haokun Zhu , Ruofei Bai , Ding Zhao , Zirui Li
show 4 more authors
Wenshan Wang Wei-Yun Yau Ji Zhang Chen Lv
This is my paper

Pith reviewed 2026-06-27 19:53 UTC · model grok-4.3

classification 💻 cs.RO
keywords object navigationhuman demonstrationsimitation learningvision-language modelsfrontier explorationspatial memoryembodiment transfer
0
0 comments X

The pith

IntentNav extracts high-level search intent from human demonstrations by labeling frontiers and trains a VLM to select among spatial-visual candidates, reaching state-of-the-art object navigation that transfers zero-shot across robot bodies

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets out to demonstrate that human navigation demonstrations contain transferable high-level intent that can be recovered by looking ahead to label which frontier best explains the demonstrator's future path. This labeled intent is then used to supervise a vision-language model that chooses among grounded candidates in a combined bird's-eye-view and egocentric memory representation. A reader would care because object navigation under partial observability remains a central unsolved robotics problem, and learning directly from human data could sidestep the expense of collecting robot-specific trajectories. The method therefore couples spatial memory of explored regions with semantic visual cues to produce exploration that avoids redundant revisits while focusing on promising areas.

Core claim

IntentNav introduces Frontier-based Human-Intent Labeling that looks ahead in human demonstrations to assign each action sequence to the frontier that best accounts for the demonstrator's future search direction. It constructs a spatial-visual candidate space in which BEV memory records explored regions, unexplored frontiers and trajectory history while egocentric visual memory supplies semantic information for each candidate; a VLM policy is then trained on these candidates under an Intent-Aligned Objective that favors consistent, human-like selections. The resulting system attains state-of-the-art success rates on the MP3D, HM3D-v1 and HM3D-v2 ObjectNav benchmarks and its candidate-level i

What carries the argument

Frontier-based Human-Intent Labeling, which looks ahead in demonstrations to assign search intent to the frontier that explains future direction and supplies these labels to a spatial-visual candidate space for VLM policy training.

If this is right

  • State-of-the-art performance on the MP3D, HM3D-v1 and HM3D-v2 ObjectNav benchmarks.
  • The candidate-level navigation interface transfers zero-shot to wheeled, quadruped, and humanoid robots without further VLM fine-tuning.
  • Exploration avoids redundant revisits by maintaining spatial memory of explored regions while using visual cues to probe promising frontiers.
  • Imitation from human demonstrations produces policies whose high-level choices remain consistent across different robot bodies.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The separation of intent labeling from low-level control may let the same trained policy serve new robot platforms by swapping only the low-level executor.
  • Frontier-based intent extraction could extend to other partially observable search tasks such as mapping or inspection without requiring embodiment-specific retraining.
  • Grounding VLM decisions in an explicit spatial-visual candidate space may reduce hallucinated exploration paths compared with purely visual end-to-end policies.

Load-bearing premise

Frontier labels derived from looking ahead in human demonstrations accurately capture transferable high-level search intent that generalizes from human data to robot execution across different embodiments.

What would settle it

A controlled experiment that replaces the human-intent frontier labels with random or heuristic labels during training and measures whether the resulting VLM policy still achieves the reported state-of-the-art success rates on HM3D-v2.

Figures

Figures reproduced from arXiv: 2606.08029 by Chen Lv, Ding Zhao, Haokun Zhu, Ji Zhang, Maonan Wang, Muyi Bao, Ruofei Bai, Wei-Yun Yau, Wenshan Wang, Yuxin Cai, Zirui Li, Zongtai Li.

Figure 1
Figure 1. Figure 1: IntentNav learns spatial-visual ObjectNav from human demonstrations. It grounds frontier and target decisions in a unified BEV space, yielding directed search behavior, strong bench￾mark performance, and transfer across wheeled, quadruped, and humanoid robots. Abstract: Object navigation requires a robot to search for an unobserved target in an unknown environment by deciding where to explore next under pa… view at source ↗
Figure 2
Figure 2. Figure 2: IntentNav overview. Left: ObjectNav is formulated as BEV candidate-level waypoint selection, with Frontier-based Human-Intent Labeling recovering supervision from low-level human rollouts. Right: The VLM policy combines the BEV decision space, candidate-aligned visual mem￾ories, and agent-centric geometry, then predicts a reserved candidate-ID token as the next waypoint. executed at least six move-forward … view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative search behavior examples. IntentNav expands unexplored regions through candidate waypoints, scans locally around informative viewpoints, and avoids redundant revisits. We also evaluated 300 human demonstration episodes on proposed metrics. Humans achieve the lowest RRR and the highest PMR, reflecting decisive search with purposeful probing and few redun￾dant revisits. However, their ECR is lowe… view at source ↗
Figure 4
Figure 4. Figure 4: Wheeled robot, object goal: cabinet. The agent navigates through a corridor lined with metal lockers. The BEV map (bottom-left) reveals that the agent initially makes a brief exploratory detour into a side passage before promptly reversing and committing to the main corridor. This early backtracking illustrates the policy’s ability to quickly abandon unproductive directions and refocus exploration toward t… view at source ↗
Figure 5
Figure 5. Figure 5: Wheeled robot, object goal: couch. The agent begins in a cluttered lounge area sur￾rounded by a ping-pong table, a foosball table, and modular seating. The BEV map (bottom-left) shows a dense initial environment with many obstacles, yet the agent efficiently sweeps through the space, exits the cluttered zone, and redirects exploration toward more promising frontiers. The pol￾icy then correctly identifies t… view at source ↗
Figure 6
Figure 6. Figure 6: Unitree Go2 quadruped, object goal: potted plant. The SLAM map (bottom-left) re￾veals a multi-room exploration pattern: the agent first sweeps the outer office area, and after finding no target there, proceeds into the inner room where the potted plant is located. This episode demon￾strates the policy’s ability to conduct structured, room-by-room search in a multi-room environment [PITH_FULL_IMAGE:figures… view at source ↗
Figure 7
Figure 7. Figure 7: Unitree Go2 quadruped, object goal: trashcan. The trajectory on the SLAM map (bottom-left) is notably long, spanning a large portion of the building across multiple corridors. Despite the extensive traversal distance, the agent’s decisions are always based on the local BEV crop, which only considers nearby frontiers. This episode highlights that the local BEV abstraction scales well to large environments: … view at source ↗
Figure 8
Figure 8. Figure 8: Unitree G1 humanoid, object goal: printer. The humanoid navigates a densely furnished office workspace. The trajectory on the SLAM map (bottom-left) shows the agent systematically probing multiple side corridors between workstations before reversing each time, demonstrating a methodical search pattern. After exhausting several dead-end branches, the policy commits to the correct direction and locates the p… view at source ↗
Figure 9
Figure 9. Figure 9: Unitree G1 humanoid, object goal: refrigerator. The left panel shows the final frame with the refrigerator visible behind the robot after approach. The trajectory on the SLAM map (top-right) shows the agent first explored multiple directions in the cubicle area before committing to the target. This behavior resembles human-like search: the agent systematically surveys the environment and immediately commit… view at source ↗
read the original abstract

Object navigation requires a robot to search for an unobserved target in an unknown environment by deciding where to explore next under partial observability. Effective search resembles human-like exploration: selectively probing visually promising frontiers while relying on spatial memory to avoid redundant revisits. We propose IntentNav, a spatial-visual imitation framework that learns human-like ObjectNav policies from human demonstrations. To infer high-level search intent from low-level human actions, we introduce Frontier-based Human-Intent Labeling, which looks ahead in human demonstrations and labels the frontier that best explains the demonstrator's future search direction. We construct a spatial-visual candidate space, where BEV memory tracks explored regions, unexplored frontiers, and trajectory history, while egocentric visual memory provides semantic cues for each candidate. A VLM policy is trained to select among these grounded candidates, using Intent-Aligned Objective to encourage consistent and human-like exploration. IntentNav achieves state-of-the-art performance on the MP3D, HM3D-v1 and HM3D-v2 ObjectNav benchmarks. The proposed candidate-level navigation interface transfers zero-shot to wheeled, quadruped, and humanoid robots without further VLM fine-tuning. \href{https://anonymous.4open.science/w/IntentNav/}{Project page}.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper introduces IntentNav, a spatial-visual imitation learning framework for object navigation that infers high-level search intent from human demonstrations via Frontier-based Human-Intent Labeling (looking ahead in trajectories to label explanatory frontiers). It constructs a candidate space combining BEV memory (for explored regions, frontiers, and history) with egocentric visual memory (for semantic cues), then trains a VLM policy to select among candidates using an Intent-Aligned Objective. The manuscript claims state-of-the-art performance on the MP3D, HM3D-v1, and HM3D-v2 ObjectNav benchmarks along with zero-shot transfer of the candidate-level interface to wheeled, quadruped, and humanoid robots without further VLM fine-tuning.

Significance. If the results hold, particularly the zero-shot embodiment transfer, the work would be significant for robot navigation by showing how human demonstration data can yield VLM policies that generalize across platforms without robot-specific retraining. The frontier-labeling procedure and structured spatial-visual candidate space offer a concrete mechanism for grounding high-level intent, which could lower data collection costs and improve sim-to-real transfer in partial-observability search tasks.

major comments (2)
  1. [Abstract] Abstract: The assertion of state-of-the-art performance on MP3D, HM3D-v1 and HM3D-v2 is presented without any quantitative metrics, baseline comparisons, ablation results, or evaluation protocol details, which is load-bearing for the central claim of superiority over prior methods.
  2. [Abstract] Abstract: The zero-shot transfer claim to wheeled, quadruped, and humanoid robots without VLM fine-tuning relies on the untested assumption that the spatial-visual candidate space (BEV + egocentric) plus Intent-Aligned Objective produces decisions invariant to kinematics, sensor placement, and reachable frontiers; no quantitative ablation or measurement of the embodiment gap or viewpoint shift is supplied, directly undermining validation of the transfer result.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for highlighting issues in the abstract that affect the clarity of our central claims. We agree that the abstract requires strengthening with concrete metrics and will revise it accordingly. We address each comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The assertion of state-of-the-art performance on MP3D, HM3D-v1 and HM3D-v2 is presented without any quantitative metrics, baseline comparisons, ablation results, or evaluation protocol details, which is load-bearing for the central claim of superiority over prior methods.

    Authors: We agree that the abstract should include supporting quantitative evidence. In the revised manuscript we will add concise performance numbers (e.g., success-rate gains over the strongest baselines on each benchmark) and a brief reference to the standard ObjectNav evaluation protocol and main baselines. Full tables, ablations, and protocol details remain in the experiments section. revision: yes

  2. Referee: [Abstract] Abstract: The zero-shot transfer claim to wheeled, quadruped, and humanoid robots without VLM fine-tuning relies on the untested assumption that the spatial-visual candidate space (BEV + egocentric) plus Intent-Aligned Objective produces decisions invariant to kinematics, sensor placement, and reachable frontiers; no quantitative ablation or measurement of the embodiment gap or viewpoint shift is supplied, directly undermining validation of the transfer result.

    Authors: The manuscript already reports successful zero-shot deployments on the three platforms using the same candidate interface and policy. We acknowledge that explicit quantitative ablations isolating the embodiment gap or viewpoint shift are not present. The revision will (1) reference the transfer experiments in the abstract and (2) add a short discussion of the design choices intended to promote embodiment invariance. We do not plan new experiments for this revision. revision: partial

Circularity Check

0 steps flagged

No circularity in derivation chain; method is standard imitation learning with empirical claims

full rationale

The paper presents a spatial-visual imitation learning framework using Frontier-based Human-Intent Labeling on human demonstrations to train a VLM policy for candidate selection. No equations, derivations, or parameter-fitting steps are described that reduce by construction to their own inputs. Claims of SOTA performance and zero-shot transfer are empirical assertions, not self-referential definitions or predictions forced by fitted parameters. No self-citation load-bearing or uniqueness theorems are invoked in the provided text. The derivation chain is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No technical details, equations, or methods sections available in the abstract; ledger left empty.

pith-pipeline@v0.9.1-grok · 5789 in / 1039 out tokens · 20118 ms · 2026-06-27T19:53:24.227166+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 23 canonical work pages · 6 internal anchors

  1. [1]

    Batra, A

    D. Batra, A. Gokaslan, A. Kembhavi, O. Maksymets, R. Mottaghi, M. Savva, A. Toshev, and E. Wijmans. ObjectNav Revisited: On Evaluation of Embodied Agents Navigating to Objects. InarXiv:2006.13171, 2020

    work page arXiv 2006

  2. [2]

    On Evaluation of Embodied Navigation Agents

    P. Anderson, A. Chang, D. S. Chaplot, A. Dosovitskiy, S. Gupta, V . Koltun, J. Kosecka, J. Ma- lik, R. Mottaghi, M. Savva, et al. On evaluation of embodied navigation agents.arXiv preprint arXiv:1807.06757, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  3. [3]

    J. Sun, J. Wu, Z. Ji, and Y .-K. Lai. A survey of object goal navigation.IEEE Transactions on Automation Science and Engineering, 2024

    2024

  4. [4]

    K. Zhou, K. Zheng, C. Pryor, Y . Shen, H. Jin, L. Getoor, and X. E. Wang. Esc: Exploration with soft commonsense constraints for zero-shot object navigation, 2023. URLhttps:// arxiv.org/abs/2301.13166

    work page arXiv 2023

  5. [5]

    Yokoyama, S

    N. Yokoyama, S. Ha, D. Batra, J. Wang, and B. Bucher. Vlfm: Vision-language frontier maps for zero-shot semantic navigation. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 42–48. IEEE, 2024

    2024

  6. [6]

    Kuang, H

    Y . Kuang, H. Lin, and M. Jiang. Openfmnav: Towards open-set zero-shot object navigation via vision-language foundation models. InFindings of the Association for Computational Linguistics: NAACL 2024, pages 338–351, 2024

    2024

  7. [7]

    H. Yin, X. Xu, Z. Wu, J. Zhou, and J. Lu. Sg-nav: Online 3d scene graph prompting for llm- based zero-shot object navigation.Advances in neural information processing systems, 37: 5285–5307, 2024

    2024

  8. [8]

    Y . Cao, J. Zhang, Z. Yu, S. Liu, Z. Qin, Q. Zou, B. Du, and K. Xu. Cognav: Cognitive process modeling for object goal navigation with llms, 2025. URLhttps://arxiv.org/abs/ 2412.10439

    work page arXiv 2025

  9. [9]

    H. Zhu, Z. Li, Z. Liu, W. Wang, J. Zhang, J. Francis, and J. Oh. Strive: Structured representation integrating vlm reasoning for efficient object navigation.arXiv preprint arXiv:2505.06729, 2025

    work page arXiv 2025

  10. [10]

    Zhang, F

    Z. Zhang, F. Hu, J. Lee, F. Shi, P. Kordjamshidi, J. Chai, and Z. Ma. Do vision-language models represent space and how? evaluating spatial frame of reference under ambiguities. arXiv preprint arXiv:2410.17385, 2024

    work page arXiv 2024

  11. [11]

    B. Chen, Z. Xu, S. Kirmani, B. Ichter, D. Sadigh, L. Guibas, and F. Xia. Spatialvlm: Endowing vision-language models with spatial reasoning capabilities. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14455–14465, 2024

    2024

  12. [12]

    Y .-H. H. Tsai, V . Dhar, J. Li, B. Zhang, and J. Zhang. Multimodal large language model for visual navigation, 2023. URLhttps://arxiv.org/abs/2310.08669

    work page arXiv 2023

  13. [13]

    L. Li, J. Zhao, Y . Xie, X. Tan, and X. Li. Compassnav: Steering from path imitation to decision understanding in navigation.arXiv preprint arXiv:2510.10154, 2025

    work page arXiv 2025

  14. [14]

    Z. Wang, H. Fang, S. Wang, Y . Luo, H. Dong, W. Li, and Y . Gan. Hydra-nav: Object navigation via adaptive dual-process reasoning.arXiv preprint arXiv:2602.09972, 2026. 9

    work page arXiv 2026

  15. [15]

    NaVid: Video-based VLM Plans the Next Step for Vision-and-Language Navigation

    J. Zhang, K. Wang, R. Xu, G. Zhou, Y . Hong, X. Fang, Q. Wu, Z. Zhang, and H. Wang. Navid: Video-based vlm plans the next step for vision-and-language navigation, 2024. URL https://arxiv.org/abs/2402.15852

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  16. [16]

    Uni-NaVid: A Video-based Vision-Language-Action Model for Unifying Embodied Navigation Tasks

    J. Zhang, K. Wang, S. Wang, M. Li, H. Liu, S. Wei, Z. Wang, Z. Zhang, and H. Wang. Uni- navid: A video-based vision-language-action model for unifying embodied navigation tasks. arXiv preprint arXiv:2412.06224, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  17. [17]

    X. Xue, J. Hu, M. Luo, X. Shichao, J. Chen, Z. Xie, Q. Kuichen, G. Wei, M. Xu, and Z. Chu. Omninav: A unified framework for prospective exploration and visual-language navigation. arXiv preprint arXiv:2509.25687, 2025

    work page arXiv 2025

  18. [18]

    H. Zhu, Z. Li, Z. Liu, K. Guo, Z. Lin, Y . Cai, G. Chen, C. Lv, W. Wang, J. Oh, et al. Sysnav: Multi-level systematic cooperation enables real-world, cross-embodiment object navigation. arXiv preprint arXiv:2603.06914, 2026

    work page arXiv 2026

  19. [19]

    P. Wu, Y . Mu, B. Wu, Y . Hou, J. Ma, S. Zhang, and C. Liu. V oronav: V oronoi-based zero-shot object navigation with large language model.arXiv preprint arXiv:2401.02695, 2024

    work page arXiv 2024

  20. [20]

    Majumdar, G

    A. Majumdar, G. Aggarwal, B. Devnani, J. Hoffman, and D. Batra. Zson: Zero-shot object- goal navigation using multimodal goal embeddings.Advances in Neural Information Process- ing Systems, 35:32340–32352, 2022

    2022

  21. [21]

    B. Yu, H. Kasaei, and M. Cao. L3mvn: Leveraging large language models for visual target navigation. In2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3554–3560. IEEE, 2023

    2023

  22. [22]

    Zhang, Y

    M. Zhang, Y . Du, C. Wu, J. Zhou, Z. Qi, J. Ma, and B. Zhou. Apexnav: An adaptive exploration strategy for zero-shot object navigation with target-centric semantic fusion.IEEE Robotics and Automation Letters, 2025

    2025

  23. [23]

    Zhong, C

    L. Zhong, C. Gao, Z. Ding, Y . Liao, H. Ma, S. Zhang, X. Zhou, and S. Liu. Topv-nav: Un- locking the top-view spatial reasoning potential of mllm for zero-shot object navigation, 2025. URLhttps://arxiv.org/abs/2411.16425

    work page arXiv 2025

  24. [24]

    Ramrakhya, E

    R. Ramrakhya, E. Undersander, D. Batra, and A. Das. Habitat-web: Learning embodied object- search strategies from human demonstrations at scale. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 5173–5183, 2022

    2022

  25. [25]

    Matterport3D: Learning from RGB-D Data in Indoor Environments

    A. Chang, A. Dai, T. Funkhouser, M. Halber, M. Niessner, M. Savva, S. Song, A. Zeng, and Y . Zhang. Matterport3d: Learning from rgb-d data in indoor environments.arXiv preprint arXiv:1709.06158, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  26. [26]

    X. Puig, E. Undersander, A. Szot, M. D. Cote, T.-Y . Yang, R. Partsey, R. Desai, A. W. Clegg, M. Hlavac, S. Y . Min, et al. Habitat 3.0: A co-habitat for humans, avatars and robots.arXiv preprint arXiv:2310.13724, 2023

    work page arXiv 2023

  27. [27]

    Z. Chen, W. Wang, Y . Cao, Y . Liu, Z. Gao, E. Cui, J. Zhu, S. Ye, H. Tian, Z. Liu, et al. Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models.arXiv preprint arXiv:2504.10479, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  28. [28]

    Yadav, S

    K. Yadav, S. K. Ramakrishnan, J. Turner, A. Gokaslan, O. Maksymets, R. Jain, R. Ramrakhya, A. X. Chang, A. Clegg, M. Savva, E. Undersander, D. S. Chaplot, and D. Batra. Habitat challenge 2022.https://aihabitat.org/challenge/2022/, 2022

    2022

  29. [29]

    Yadav, R

    K. Yadav, R. Ramrakhya, S. K. Ramakrishnan, T. Gervet, J. Turner, A. Gokaslan, N. Maestre, A. X. Chang, D. Batra, M. Savva, et al. Habitat-matterport 3d semantics dataset. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4927–4936, 2023. 10

    2023

  30. [30]

    Zhang, Q

    L. Zhang, Q. Zhang, H. Wang, E. Xiao, Z. Jiang, H. Chen, and R. Xu. Trihelper: Zero- shot object navigation with dynamic assistance, 2024. URLhttps://arxiv.org/abs/ 2403.15223

    work page arXiv 2024

  31. [31]

    Z. Zhou, Y . Hu, L. Zhang, Z. Li, and S. Chen. Beliefmapnav: 3d voxel-based belief map for zero-shot object navigation.arXiv preprint arXiv:2506.06487, 2025

    work page arXiv 2025

  32. [32]

    D. Nie, X. Guo, Y . Duan, R. Zhang, and L. Chen. Wmnav: Integrating vision-language models into world models for object goal navigation, 2025. URLhttps://arxiv.org/abs/ 2503.02247

    work page arXiv 2025

  33. [33]

    C. Peng, Z. Zhang, C. Chi, X. Wei, Y . Zhang, H. Wang, P. Wang, Z. Wang, J. Liu, and S. Zhang. Pigeon: Vlm-driven object navigation via points of interest selection.arXiv preprint arXiv:2511.13207, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  34. [34]

    Wijmans, A

    E. Wijmans, A. Kadian, A. Morcos, S. Lee, I. Essa, D. Parikh, M. Savva, and D. Batra. Dd-ppo: Learning near-perfect pointgoal navigators from 2.5 billion frames.arXiv preprint arXiv:1911.00357, 2019

    work page arXiv 1911

  35. [35]

    W. Cai, S. Huang, G. Cheng, Y . Long, P. Gao, C. Sun, and H. Dong. Bridging zero-shot object navigation and foundation models through pixel-guided navigation skill. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 5228–5234. IEEE, 2024

    2024

  36. [36]

    S. K. Ramakrishnan, D. S. Chaplot, Z. Al-Halah, J. Malik, and K. Grauman. Poni: Po- tential functions for objectgoal navigation with interaction-free learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18890–18900, 2022

    2022

  37. [37]

    D. S. Chaplot, D. P. Gandhi, A. Gupta, and R. R. Salakhutdinov. Object goal navigation using goal-oriented semantic exploration.Advances in Neural Information Processing Systems, 33: 4247–4258, 2020

    2020

  38. [38]

    H. Yin, X. Xu, L. Zhao, Z. Wang, J. Zhou, and J. Lu. Unigoal: Towards universal zero-shot goal-oriented navigation. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 19057–19066, 2025. 11 Appendix A BEV Map Construction from RGB-D Observations This section describes the BEV mapping procedure used to construct the spatial input...

    2025