Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

Learning Camera Movement Control from Real-World Drone Videos

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read An AI cameraman can learn skilled drone camera movement from 99,003 trajectories automatically extracted from real YouTube drone videos, and beats a robotics baseline on user preference, collision rate, and smoothness in simulation.

desk verdict A useful dataset and a clean architecture, but the scale-normalization-to-execution gap leaves the headline numbers under-specified. read the letter →

arxiv 2412.09620 v1 pith:FDIXARIV submitted 2024-12-12 cs.CV cs.RO

Yunzhong Hou , Liang Zheng , Philip Torr This is my paper · ORCID
classification cs.CVcs.RO
keywords dronevideographycameramovementcontrolautoregressivetransformerstructure-from-motionKalmanfilterbehaviorcloningreal-worlddataset3Dtrajectory
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper aims to show that camera movement control for drone videography can be learned at scale from real-world online videos, without hand-coded cinematography rules or expensive expert teleoperation. To do this, the authors automatically extract 99,003 three-dimensional camera trajectories from YouTube drone footage using structure-from-motion and a Kalman filter, then train DVGFormer, an autoregressive transformer that predicts the next camera motion from all past images, poses, and motions. In simulation over 38 synthetic natural scenes and 7 real city 3D scans, the model learns challenging moves such as navigating through obstacles, flying low to increase perceived speed, and orbiting towers and buildings, and it records higher user preference, a lower collision rate, and smoother motion than an RT-1-inspired baseline. The significance is that this points to a scalable path to AI videography that does not require pixel generation or heuristic objective design.

What carries the argument

The load-bearing machinery is the automatic trajectory-extraction pipeline that builds DroneMotion-99k together with the DVGFormer architecture that consumes it. The pipeline splits scraped YouTube videos into shots with shot-change detection, recovers per-frame camera poses with Colmap structure-from-motion, connects consecutive poses into three-dimensional paths, and uses an Unscented Kalman Filter plus a labeled threshold to discard reconstructions whose camera locations jump between neighboring frames; a scale normalization divides each scene by the average inter-frame camera distance, assuming drone speed is stable across videos. DVGFormer is a GPT-2-style autoregressive transformer of about 40 million parameters that tokenizes camera pose and motion with MLPs, image patches with DINOv2 features, and monocular depth from Depth-Anything, breaks each predicted action into five sub-steps, and predicts the next camera motion from up to 10 seconds of past frames using bi-level positional embeddings.

What would settle it

Record drone flights whose true three-dimensional trajectories are known from onboard GPS/IMU telemetry, deliberately covering a wide range of flight speeds, run the paper's Colmap-and-normalization pipeline on that footage, and compare each normalized camera path to the true flight speed; if the normalized trajectories do not track the true speeds, the speed-stability assumption fails and the learned camera motions would be miscalibrated.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a long-horizon behavior-cloning model can acquire sophisticated drone cinematography skills purely from automatically recovered real-world camera paths. The authors train DVGFormer on the DroneMotion-99k dataset, which contains 99,003 trajectories totaling more than 180 hours of footage, and report that compared with an RT-1-inspired baseline it achieves significantly better user preference (70.5% versus 29.5%), a lower collision rate (15.2% versus 33.7%), and smoother motion, measured as smaller maximum relative changes in velocity and angular velocity, across 184 simulated videos covering 38 natural scenes and 7 real city scans. Qualitatively, they observe behaviors such as obstacle navigation, low-altitude flyovers, building orbits, and even transfer to underwater scenes, all without hand-written cinematography heuristics.

Load-bearing premise

The whole dataset is normalized by the average inter-frame camera distance under the assumption that drone flight speed is roughly stable across different videos; if real speeds vary substantially across the scraped footage, the recovered trajectories carry inconsistent scale and the model would learn distorted motion magnitudes and speeds.

Editorial extensions

If this is right

  • Videography control for drones can be built from already-existing online video at low annotation cost, moving the data bottleneck from expert teleoperation to automatic reconstruction and filtering.
  • Because the model outputs executable three-dimensional trajectories rather than pixels, the same trained system can be deployed in any renderable 3D scene, including ones unlike its training footage.
  • The action-chunking design and long context window let the model continue trajectories beyond its training length, producing smooth 20-second videos without post-processing.
  • The <Cond> token provides a stochastic control handle: the same starting frame can yield multiple distinct but feasible camera paths.
  • The comparison with the RT-1-inspired baseline indicates that adding camera pose and motion tokens plus a longer temporal receptive field is what separates smooth videography from a generalist behavior-cloning model.

Reading between the lines

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

  • A natural test the paper leaves implicit is to validate the extraction pipeline on footage with known UAV telemetry; if flight speeds vary widely in practice, the normalization would distort recovered trajectory magnitudes and speeds.
  • The same automatic-trajectory recipe could transfer to other camera platforms such as handheld gimbals, dashcams, or underwater ROVs, potentially yielding a general camera-movement prior for videography.
  • The reported crash rate suggests a learned movement prior alone is not yet safe for real-world flight; adding an explicit collision-avoidance term or sequence-level 3D perception at inference time is a concrete next step the paper itself flags.
  • The 99k real trajectories could serve as a camera-motion prior for generative video models that condition on camera paths, narrowing the search space without pixel-space supervision.
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

4 major / 5 minor

Summary. The paper introduces DroneMotion-99k, a dataset of 99,003 camera trajectories recovered from YouTube drone videos via Colmap reconstruction and Kalman-filter-based quality filtering, and DVGFormer, an autoregressive transformer that predicts 6-DoF camera motion from past frames, camera poses, and depth estimates. The system is evaluated in Blender on 38 synthetic natural scenes and 7 Google Earth city scans against an RT-1-inspired baseline, reporting higher user preference, lower collision rate, and smoother motion. The core idea is scalable and avoids hand-crafted cinematography heuristics, but the evaluation is thin and the scale-normalization step is load-bearing for the main claims.

Significance. If the scale ambiguity is resolved, this is a valuable contribution to AI videography: it provides a large real-world trajectory dataset, an automatic annotation pipeline, and a model that exploits long-horizon context. The authors state that data and code will be released, which would aid reproducibility. However, the current evidence for the central claim rests on a single baseline, a simulation-only evaluation without error bars or significance tests, and an unverified assumption about uniform drone speed during dataset normalization. The authors' own limitation section acknowledges the scale issue and the undesirable crash rate, which is honest but also indicates that the main claims are not yet fully supported.

major comments (4)
  1. [Sec. 3.2, Sec. 4.1, Sec. 5.2] The scale-normalization assumption is load-bearing and is never resolved end-to-end. Section 3.2 normalizes each 3D reconstruction by the average inter-frame camera distance, explicitly assuming stable drone speed across videos, and Appendix A further discards clips whose maximum speed exceeds 3 times the average. Section 4.1 says predicted motions are normalized according to dataset statistics, but the paper never states how these normalized outputs are mapped to metric units when executed in the Blender evaluation platform, where city scans have real metric extent (e.g., roughly 1 km regions). Without this mapping, the collision rate in Table 2 and the qualitative claims about low-altitude flight and obstacle navigation are not well-defined: direct execution in normalized units makes absolute trajectory scale arbitrary, while any fixed de-normalization constant inherits the unverified uniform-speed assumption. Please report the exact inference-time scale mapping and justify it, or evaluate how collision rate and user preference vary under a plausible range of scale factors.
  2. [Sec. 5.4, Table 2] The quantitative evidence for the central claim is not statistically grounded. Table 2 reports user preference, collision rate, and smoothness on 184 videos, but no confidence intervals, number of human raters, or significance tests are provided, so the word 'significantly' in the Abstract and Section 5.4 is not supported. Moreover, the only comparison is an RT-1-inspired baseline; existing cinematography methods are set aside in Section 5.3 because they require human actors, but a learning-based method such as Jiang et al. [38] or a simple heuristic controller could be adapted to the same 3D scenes. Please add variance estimates, a paired significance test for preference and collision rate, and at least one additional baseline, or soften the corresponding claims.
  3. [Abstract, Sec. 5.5] The Abstract claims that the system 'effectively learns' to navigate through obstacles, maintain low altitude, and orbit towers and buildings. These behaviors appear only as qualitative examples in Fig. 1 and Fig. 5 and are not quantified. Define and report metrics such as distance to the nearest scene mesh, altitude above ground, and orbit coverage for both DVGFormer and the baseline, and compare against a random-trajectory control. Without such metrics, the distinguishing behaviors claimed in the Abstract are not testable and the central qualitative claim is under-supported.
  4. [Sec. 3.3, Appendix A] The data-filtering pipeline relies on two thresholds whose sensitivity is not analyzed: the Kalman-filter distance threshold of 0.2 selected on roughly 1,000 labeled reconstructions, and the rule discarding clips whose maximum speed exceeds 3 times the average speed. The paper does not report whether the labeled reconstructions were split to avoid overfitting the threshold, nor how downstream model performance changes when these thresholds are varied. Please report this sensitivity or justify the choices with cross-validation, since these thresholds determine the final dataset composition.
minor comments (5)
  1. [Sec. 6] The section heading contains a typo: 'Limitatinons' should be 'Limitations'.
  2. [Appendix C] In the generalization study description, 'we the under water setting in InfiniGen' appears to be missing a verb, likely 'we use the under water setting in InfiniGen'.
  3. [Sec. 5.2] The phrase 'natural and civic scenes' should probably be 'natural and city scenes' for clarity.
  4. [Table 2] The caption should state how the user preference percentage is computed (e.g., pairwise votes, fraction of videos preferred) and report the number of human raters.
  5. [Sec. 4.1] Minor grammatical issue: 'the models predicts' should be 'the model predicts'; also, the sentence 'we break the action at into N steps' is missing the subscript t in the running text, which can confuse readers.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the derivation chain is self-contained—real trajectories are reconstructed and filtered, the model is trained by behavior cloning, and evaluation uses held-out rendered scenes with user preference, collision, and smoothness metrics that are not encoded in the training targets.

full rationale

The paper's central claim—that DVGFormer trained on DroneMotion-99k produces preferred, low-collision, smooth trajectories—is not baked into its inputs. Dataset construction (Sec. 3) extracts camera poses from YouTube videos with Colmap and filters low-quality reconstructions using a Kalman-filter deviation threshold selected on roughly 1k manually labeled reconstructions; this selection concerns reconstruction correctness, not the downstream metrics of user preference or collision. The model (Sec. 4) is an autoregressive transformer trained with an L1 loss to regress 6-DoF camera motions from past poses, images, and depth; no term in the loss encodes the evaluation criteria. The RT-1 inspired baseline is a genuine comparator with the same image features but a shorter horizon and no pose/motion tokens. Evaluation (Sec. 5.4) is on held-out Blender-rendered natural scenes and city scans with matched initial poses, and user preference, collision rate, and smoothness are measured from rendered videos, not from training statistics. The scale-ambiguity caveat in Sec. 3.2 and the discussion of the uniform-drone-speed assumption are limitations about metric interpretation and generalization, not circular reasoning: normalizing each clip by average inter-frame distance does not determine the model's output preferences, and the paper explicitly acknowledges the scale issue and reports unsuccessful attempts with metric depth. There are no load-bearing self-citations and no renamed known result; the comparison to RT-1 is an external baseline, not an imported conclusion. Therefore the circularity score is 0.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a data processing pipeline with two hand-set thresholds and several domain assumptions about video quality, scale normalization, and simulation validity. No new physical entities are introduced.

free parameters (2)
  • Kalman filter distance threshold = 0.2
    Chosen from a ROC/precision-recall analysis on about 1k manually labeled reconstructions to separate correct and incorrect reconstructions (Sec. 3.3, Fig. 3). Affects which trajectories enter the 99k dataset.
  • Maximum speed multiplier = 3x average speed
    Reconstructions whose maximum camera speed exceeds 3 times the average are discarded (Appendix A). Hand-chosen threshold that changes the dataset distribution.
assumptions (5)
  • domain assumption The drone's moving speed is roughly stable across different videos, so normalizing each reconstruction by average inter-frame distance gives consistent scale.
    Invoked in Sec. 3.2 to resolve scale ambiguity of SfM reconstructions; load-bearing for dataset correctness.
  • domain assumption Camera trajectories extracted from online 'cinematic drone' videos are good demonstrations of videography.
    The dataset is scraped with keywords like 'cinematic drone footage' and filtered by reconstruction quality, but the aesthetic quality of the resulting trajectories is not independently validated (Sec. 3.1, Appendix A).
  • domain assumption Evaluation in Blender with InfiniGen synthetic scenes and Google Earth city meshes is a valid proxy for real-world drone cinematography.
    All quantitative results are from simulation (Sec. 5.2); the paper notes the approach still has an undesirable crash rate for real-world piloting (Sec. 6).
  • domain assumption Monocular depth from Depth-Anything provides sufficient 3D awareness for collision avoidance in novel scenes.
    DVGFormer uses per-frame monocular depth as the only geometric signal beyond the camera path (Sec. 4.2); the paper acknowledges SLAM might be better at inference (Sec. 6).
  • standard math Structure-from-Motion (Colmap) and Unscented Kalman Filter work as described in their citations.
    The data pipeline relies on these off-the-shelf tools; the paper assumes their outputs are reliable enough for trajectory extraction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Camera Movement Control from Real-World Drone Videos." pith.science (2026). https://pith.science/paper/FDIXARIV

@misc{pith2026241209620,
  author       = {Pith},
  title        = {Pith review of: Learning Camera Movement Control from Real-World Drone Videos},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FDIXARIV}},
  note         = {Machine review of arXiv:2412.09620}
}
read the original abstract

This study seeks to automate camera movement control for filming existing subjects into attractive videos, contrasting with the creation of non-existent content by directly generating the pixels. We select drone videos as our test case due to their rich and challenging motion patterns, distinctive viewing angles, and precise controls. Existing AI videography methods struggle with limited appearance diversity in simulation training, high costs of recording expert operations, and difficulties in designing heuristic-based goals to cover all scenarios. To avoid these issues, we propose a scalable method that involves collecting real-world training data to improve diversity, extracting camera trajectories automatically to minimize annotation costs, and training an effective architecture that does not rely on heuristics. Specifically, we collect 99k high-quality trajectories by running 3D reconstruction on online videos, connecting camera poses from consecutive frames to formulate 3D camera paths, and using Kalman filter to identify and remove low-quality data. Moreover, we introduce DVGFormer, an auto-regressive transformer that leverages the camera path and images from all past frames to predict camera movement in the next frame. We evaluate our system across 38 synthetic natural scenes and 7 real city 3D scans. We show that our system effectively learns to perform challenging camera movements such as navigating through obstacles, maintaining low altitude to increase perceived speed, and orbiting towers and buildings, which are very useful for recording high-quality videos. Data and code are available at dvgformer.github.io.

Figures

Figures reproduced from arXiv: 2412.09620 by the authors.

Figure 1
Figure 1. Examples of recorded videos from our AI cameraman. Instead of generating [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Data collection pipeline. Top left: For scraped YouTube videos, we run shot change detection [18] to split the videos into clips of individual scene. Top right: We then use Colmap [55] to reconstruct the 3D scene and recover camera poses from video frames. Bottom: Finally, we connect camera poses from consecutive frames to formulate 3D camera trajectories and apply Kalman filter [63] to discard low quality reconstru… view at source ↗
Figure 3
Figure 3. Threshold selection for identifying low-quality 3D re [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Model overview of DVGFormer. To predict camera motion [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Visualization of the recorded videos. DVGFormer learns techniques like keeping the actor in frame, navigating through obstacles, [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: 3D camera trajectory comparison. Compared to the RT-1 inspired basline ( [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 8
Figure 8. Figure 8: From the same initial image (leftmost), our model can output different but feasible camera paths using different [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 10
Figure 10. Figure 10: Clip duration distribution. 86.6% of the clips produced [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Towards Understanding Camera Motions in Any Video

    cs.CV 2025-04 conditional novelty 7.0 of 10

    CameraBench is a new expert-annotated benchmark with a 50-primitive camera motion taxonomy, revealing that SfM and VLMs fail in complementary ways and that SFT on the benchmark boosts VLM performance.

  2. EPiC: Efficient Video Camera Control Learning with Precise Anchor-Video Guidance

    cs.CV 2025-05 conditional novelty 6.0 of 10

    EPiC trains a 30M-parameter visibility-aware ControlNet on mask-based anchor videos from 5,000 in-the-wild videos and 500 steps, reaching SOTA camera accuracy on RealEstate10K and MiraData.

Reference graph

Works this paper leans on

70 extracted references · 48 canonical work pages · cited by 2 Pith papers

  1. [38]

    Example-driven virtual cinematography by learning camera behaviors

    Hongda Jiang, Bin Wang, Xi Wang, Marc Christie, and Baoquan Chen. Example-driven virtual cinematography by learning camera behaviors. ACM Trans. Graph., 39(4):45,

    work page

  2. [1]

    https://earth

    Google Earth — earth.google.com. https://earth. google.com/web/. [Accessed 06-11-2024]. 6

    work page 2024

  3. [2]

    https : / / prochitecture.gumroad.com/l/blender- osm

    Blosm for Blender: Google 3D cities, OpenStreetMap, terrain — prochitecture.gumroad.com. https : / / prochitecture.gumroad.com/l/blender- osm . [Accessed 06-11-2024]. 6, 10

    work page 2024

  4. [3]

    https : / / filterpy

    FilterPy; FilterPy 1.4.4 documentation — fil- terpy.readthedocs.io. https : / / filterpy . readthedocs.io. [Accessed 18-11-2024]. 10

    work page 2024

  5. [4]

    https://sketchfab.com/

    Sketchfab - The best 3D viewer on the web — sketch- fab.com. https://sketchfab.com/ . [Accessed 20- 11-2024]. 10

    work page 2024

  6. [5]

    com/yt-dlp/yt-dlp

    GitHub - yt-dlp/yt-dlp: A feature-rich command-line au- dio/video downloader — github.com.https://github. com/yt-dlp/yt-dlp. [Accessed 18-11-2024]. 9

    work page 2024

  7. [6]

    A drone video clip dataset and its applications in automated cinematography

    Amirsaman Ashtari, Raehyuk Jung, Mingxiao Li, and Juny- ong Noh. A drone video clip dataset and its applications in automated cinematography. In Computer Graphics Forum, pages 189–203. Wiley Online Library, 2022. 2, 3, 4

    work page 2022

  8. [7]

    Reviewing intelligent cinematography: Ai re- search for camera-based video production

    Adrian Azzarelli, Nantheera Anantrasirichai, and David R Bull. Reviewing intelligent cinematography: Ai re- search for camera-based video production. arXiv preprint arXiv:2405.05039, 2024. 2, 6

    arXiv 2024

Show all 70 references
  1. [8]

    Is space-time attention all you need for video understanding? In Proceedings of the International Conference on Machine Learning (ICML), 2021

    Gedas Bertasius, Heng Wang, and Lorenzo Torresani. Is space-time attention all you need for video understanding? In Proceedings of the International Conference on Machine Learning (ICML), 2021. 8

    2021

  2. [9]

    Stable video diffusion: Scaling latent video diffusion models to large datasets

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram V oleti, Adam Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127, 2023. 2

    2023 arXiv

  3. [10]

    Film art: An introduction

    David Bordwell, Kristin Thompson, and Jeff Smith. Film art: An introduction. McGraw-Hill New York, 2010. 3

    2010

  4. [11]

    Openai gym

    Greg Brockman, Vicki Cheung, Ludwig Pettersson, Jonas Schneider, John Schulman, Jie Tang, and Wojciech Zaremba. Openai gym. CoRR, abs/1606.01540, 2016. 8

    2016 arXiv

  5. [12]

    Rt-1: Robotics transformer for real- world control at scale

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Joseph Dabis, Chelsea Finn, Keerthana Gopalakr- ishnan, Karol Hausman, Alex Herzog, Jasmine Hsu, Ju- lian Ibarz, Brian Ichter, Alex Irpan, Tomas Jackson, Sally Jesmonth, Nikhil Joshi, Ryan Julian, Dmitry Kalashniko...

    arXiv

  6. [13]

    Rt-2: Vision-language-action models transfer web knowledge to robotic control

    Anthony Brohan, Noah Brown, Justice Carbajal, Yevgen Chebotar, Xi Chen, Krzysztof Choromanski, Tianli Ding, Danny Driess, Avinava Dubey, Chelsea Finn, Pete Florence, Chuyuan Fu, Montse Gonzalez Arenas, Keerthana Gopalakr- ishnan, Kehang Han, Karol Hausman, Alex Herzog, Jas- mi...

    2023 arXiv

  7. [14]

    Video generation models as world simulators

    Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luh- man, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. Video generation models as world simulators

  8. [15]

    Ge- nie: Generative interactive environments

    Jake Bruce, Michael D Dennis, Ashley Edwards, Jack Parker-Holder, Yuge Shi, Edward Hughes, Matthew Lai, Aditi Mavalankar, Richie Steigerwald, Chris Apps, et al. Ge- nie: Generative interactive environments. InForty-first Inter- national Conference on Machine Learning, 2024. 2

    2024

  9. [16]

    Orb-slam3: An accu- rate open-source library for visual, visual–inertial, and mul- timap slam

    Carlos Campos, Richard Elvira, Juan J G ´omez Rodr´ıguez, Jos´e MM Montiel, and Juan D Tard´os. Orb-slam3: An accu- rate open-source library for visual, visual–inertial, and mul- timap slam. IEEE Transactions on Robotics , 37(6):1874– 1890, 2021. 3

    2021

  10. [17]

    Quo vadis, action recognition? a new model and the kinetics dataset

    Joao Carreira and Andrew Zisserman. Quo vadis, action recognition? a new model and the kinetics dataset. In pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6299–6308, 2017. 8

    2017

  11. [18]

    Home - PySceneDetect — scenede- tect.com

    Brandon Castellano. Home - PySceneDetect — scenede- tect.com. https://www.scenedetect.com/ . [Ac- cessed 04-11-2024]. 3, 9

    2024

  12. [19]

    Decision transformer: Reinforce- ment learning via sequence modeling

    Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Misha Laskin, Pieter Abbeel, Aravind Srini- vas, and Igor Mordatch. Decision transformer: Reinforce- ment learning via sequence modeling. Advances in neural information processing systems, 34:15084–15097, 2021. 5, 8

    2021

  13. [20]

    Videography: video media as art and culture

    Sean Cubitt. Videography: video media as art and culture . Bloomsbury Publishing, 1993. 1

    1993

  14. [21]

    Quicker, faster, darker: Changes in hollywood film over 75 years

    James E Cutting, Kaitlin L Brunick, Jordan E DeLong, Catalina Iricinschi, and Ayse Candan. Quicker, faster, darker: Changes in hollywood film over 75 years. i- Perception, 2(6):569–576, 2011. 3

    2011

  15. [22]

    Carbonell, Quoc V

    Zihang Dai, Zhilin Yang, Yiming Yang, Jaime G. Carbonell, Quoc V . Le, and Ruslan Salakhutdinov. Transformer-xl: Attentive language models beyond a fixed-length context. CoRR, abs/1901.02860, 2019. 4, 6, 7

    1901 arXiv

  16. [23]

    WATCH: Ultimate Guide to Cam- era Movement: When & How to Use Camera Move- ment

    Kyle DeGuzman. WATCH: Ultimate Guide to Cam- era Movement: When & How to Use Camera Move- ment... Explained — studiobinder.com. https://www. studiobinder . com / blog / different - types - of- camera- movements- in- film/. [Accessed 21- 10-2024]. 1

    2024

  17. [24]

    Domain-size pooling in local descriptors: Dsp-sift

    Jingming Dong and Stefano Soatto. Domain-size pooling in local descriptors: Dsp-sift. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 5097–5106, 2015. 3

    2015

  18. [25]

    An image is worth 16x16 words: Transformers for image recognition at scale, 2021

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2021

  19. [26]

    blender.org - Home of the Blender project - Free and Open 3D Creation Software — blender.org

    Blender Foundation. blender.org - Home of the Blender project - Free and Open 3D Creation Software — blender.org. https://www.blender.org/. [Accessed 06-11-2024]. 5, 10

    2024

  20. [27]

    Animatediff: Animate your personalized text- to-image diffusion models without specific tuning

    Yuwei Guo, Ceyuan Yang, Anyi Rao, Zhengyang Liang, Yaohui Wang, Yu Qiao, Maneesh Agrawala, Dahua Lin, and Bo Dai. Animatediff: Animate your personalized text- to-image diffusion models without specific tuning. arXiv preprint arXiv:2307.04725, 2023. 2

    2023 arXiv

  21. [28]

    Cameractrl: Enabling camera control for text-to-video generation, 2024

    Hao He, Yinghao Xu, Yuwei Guo, Gordon Wetzstein, Bo Dai, Hongsheng Li, and Ceyuan Yang. Cameractrl: Enabling camera control for text-to-video generation, 2024. 2

    2024

  22. [29]

    Cohen, and David H

    Li-wei He, Michael F. Cohen, and David H. Salesin. The virtual cinematographer: a paradigm for automatic real-time camera control and directing. In Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques, page 217–224, New York, NY , USA, 1996...

    1996

  23. [30]

    Gaussian error linear units (gelus)

    Dan Hendrycks and Kevin Gimpel. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415, 2016. 10

    2016 arXiv

  24. [31]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 1

    2020

  25. [32]

    Act: An autonomous drone cinematography sys- tem for action scenes

    Chong Huang, Fei Gao, Jie Pan, Zhenyu Yang, Weihao Qiu, Peng Chen, Xin Yang, Shaojie Shen, and Kwang-Ting Cheng. Act: An autonomous drone cinematography sys- tem for action scenes. In 2018 ieee international conference on robotics and automation (icra), pages 7039–7046. IEEE,

    2018

  26. [33]

    Learning to film from professional human motion videos

    Chong Huang, Chuan-En Lin, Zhenyu Yang, Yan Kong, Peng Chen, Xin Yang, and Kwang-Ting Cheng. Learning to film from professional human motion videos. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4244–4253, 2019. 2, 4

    2019

  27. [34]

    Vbench: Comprehensive bench- mark suite for video generative models

    Ziqi Huang, Yinan He, Jiashuo Yu, Fan Zhang, Chenyang Si, Yuming Jiang, Yuanhan Zhang, Tianxing Wu, Qingyang Jin, Nattapol Chanpaisit, et al. Vbench: Comprehensive bench- mark suite for video generative models. In Proceedings of the IEEE/CVF Conference on Computer Vision and P...

    2024

  28. [35]

    Offline re- inforcement learning as one big sequence modeling prob- lem

    Michael Janner, Qiyang Li, and Sergey Levine. Offline re- inforcement learning as one big sequence modeling prob- lem. In Advances in Neural Information Processing Systems,

  29. [36]

    Integrated mo- tion planner for real-time aerial videography with a drone in a dense environment

    Boseong Jeon, Yunwoo Lee, and H Jin Kim. Integrated mo- tion planner for real-time aerial videography with a drone in a dense environment. In 2020 IEEE International Confer- ence on Robotics and Automation (ICRA), pages 1243–1249. IEEE, 2020. 2

    2020

  30. [37]

    Detection-aware trajectory generation for a drone cine- matographer

    Boseong Felipe Jeon, Dongsuk Shim, and H Jin Kim. Detection-aware trajectory generation for a drone cine- matographer. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1450–1457. IEEE, 2020. 2, 8

    2020

  31. [39]

    To- wards a drone cinematographer: Guiding quadrotor cam- eras using visual composition principles

    Niels Joubert, Dan B Goldman, Floraine Berthouzoz, Mike Roberts, James A Landay, Pat Hanrahan, et al. To- wards a drone cinematographer: Guiding quadrotor cam- eras using visual composition principles. arXiv preprint arXiv:1610.01691, 2016. 2, 7

    arXiv

  32. [40]

    Openvla: An open-source vision-language-action model

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan Foster, Grace Lam, Pannag Sanketi, et al. Openvla: An open-source vision-language-action model. arXiv preprint arXiv:2406.09246, 2024. 2

    2024 arXiv

  33. [41]

    Col- laborative video diffusion: Consistent multi-video genera- tion with camera control

    Zhengfei Kuang, Shengqu Cai, Hao He, Yinghao Xu, Hong- sheng Li, Leonidas Guibas, and Gordon Wetzstein. Col- laborative video diffusion: Consistent multi-video genera- tion with camera control. arXiv preprint arXiv:2405.17414,

    arXiv

  34. [42]

    Infinitenature-zero: Learning perpetual view generation of natural scenes from single images

    Zhengqi Li, Qianqian Wang, Noah Snavely, and Angjoo Kanazawa. Infinitenature-zero: Learning perpetual view generation of natural scenes from single images. In Eu- ropean Conference on Computer Vision , pages 515–534. Springer, 2022. 2

    2022

  35. [43]

    Distinctive image features from scale- invariant keypoints

    David G Lowe. Distinctive image features from scale- invariant keypoints. International journal of computer vi- sion, 60:91–110, 2004. 3, 9

    2004

  36. [44]

    High-level multiple-uav cinematography tools for covering outdoor events

    Ioannis Mademlis, Vasileios Mygdalis, Nikos Nikolaidis, Maurizio Montagnuolo, Fulvio Negro, Alberto Messina, and Ioannis Pitas. High-level multiple-uav cinematography tools for covering outdoor events. IEEE Transactions on Broad- casting, 65(3):627–635, 2019. 2

    2019

  37. [45]

    Vision-based drone control for au- tonomous uav cinematography

    Ioannis Mademlis, Charalampos Symeonidis, Anastasios Tefas, and Ioannis Pitas. Vision-based drone control for au- tonomous uav cinematography. Multimedia Tools and Appli- cations, 83(8):25055–25083, 2024. 2

    2024

  38. [46]

    Orb-slam2: An open- source slam system for monocular, stereo, and rgb-d cam- eras

    Raul Mur-Artal and Juan D Tard ´os. Orb-slam2: An open- source slam system for monocular, stereo, and rgb-d cam- eras. IEEE transactions on robotics, 33(5):1255–1262, 2017. 2

    2017

  39. [47]

    Real-time planning for automated multi-view drone cinematography

    Tobias N ¨ageli, Lukas Meier, Alexander Domahidi, Javier Alonso-Mora, and Otmar Hilliges. Real-time planning for automated multi-view drone cinematography. ACM Trans- actions on Graphics (TOG), 36(4):1–10, 2017. 2, 7

    2017

  40. [48]

    Sg-i2v: Self-guided trajectory control in image-to-video generation

    Koichi Namekata, Sherwin Bahmani, Ziyi Wu, Yash Kant, Igor Gilitschenski, and David B Lindell. Sg-i2v: Self-guided trajectory control in image-to-video generation. arXiv preprint arXiv:2411.04989, 2024. 2

    2024 arXiv

  41. [49]

    Chatgpt: Optimizing language models for dia- logue

    OpenAI. Chatgpt: Optimizing language models for dia- logue. https://www.openai.com/chatgpt , 2022. Accessed: 2024-10-21. 2

    2022

  42. [50]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 5, 10

    2023 arXiv

  43. [51]

    Movie gen: A cast of media foundation models

    Adam Polyak, Amit Zohar, Andrew Brown, Andros Tjandra, Animesh Sinha, Ann Lee, Apoorv Vyas, Bowen Shi, Chih- Yao Ma, Ching-Yao Chuang, et al. Movie gen: A cast of media foundation models. arXiv preprint arXiv:2410.13720,

    arXiv

  44. [52]

    Cinempc: A fully autonomous drone cinematography system incorporating zoom, focus, pose, and scene composition

    Pablo Pueyo, Juan Dendarieta, Eduardo Montijano, Ana C Murillo, and Mac Schwager. Cinempc: A fully autonomous drone cinematography system incorporating zoom, focus, pose, and scene composition. IEEE Transactions on Robotics, 2024. 2

    2024

  45. [53]

    Language models are unsu- pervised multitask learners

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsu- pervised multitask learners. OpenAI blog, 1(8):9, 2019. 2, 5

    2019

  46. [54]

    Infinite photore- alistic worlds using procedural generation

    Alexander Raistrick, Lahav Lipson, Zeyu Ma, Lingjie Mei, Mingzhe Wang, Yiming Zuo, Karhan Kayan, Hongyu Wen, Beining Han, Yihan Wang, Alejandro Newell, Hei Law, Ankit Goyal, Kaiyu Yang, and Jia Deng. Infinite photore- alistic worlds using procedural generation. In Proceedings ...

    2023

  47. [55]

    Structure-from-motion revisited

    Johannes Lutz Sch ¨onberger and Jan-Michael Frahm. Structure-from-motion revisited. In Conference on Com- puter Vision and Pattern Recognition (CVPR) , 2016. 2, 3, 9

    2016

  48. [56]

    ViNT: A foundation model for visual navigation

    Dhruv Shah, Ajay Sridhar, Nitish Dashora, Kyle Stachowicz, Kevin Black, Noriaki Hirose, and Sergey Levine. ViNT: A foundation model for visual navigation. In 7th Annual Con- ference on Robot Learning, 2023. 2, 7

    2023

  49. [57]

    Airsim: High-fidelity visual and physical simulation for autonomous vehicles

    Shital Shah, Debadeepta Dey, Chris Lovett, and Ashish Kapoor. Airsim: High-fidelity visual and physical simulation for autonomous vehicles. In Field and Service Robotics: Re- sults of the 11th International Conference , pages 621–635. Springer, 2018. 2, 8

    2018

  50. [58]

    DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras

    Zachary Teed and Jia Deng. DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras. Ad- vances in neural information processing systems, 2021. 3

    2021

  51. [59]

    Online Learning — Camera Movement — acmi.net.au

    the Australian Centre for the Moving Image. Online Learning — Camera Movement — acmi.net.au. https: / / www . acmi . net . au / education / school - program- and- resources/exploring- camera- movement/. [Accessed 21-10-2024]. 1

    2024

  52. [60]

    Llama: Open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timoth´ee Lacroix, Baptiste Rozi`ere, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023. 2

    2023 arXiv

  53. [61]

    Diffusion models are real-time game engines,

    Dani Valevski, Yaniv Leviathan, Moab Arar, and Shlomi Fruchter. Diffusion models are real-time game engines,

  54. [62]

    Bridgedata v2: A dataset for robot learning at scale

    Homer Walke, Kevin Black, Abraham Lee, Moo Jin Kim, Max Du, Chongyi Zheng, Tony Zhao, Philippe Hansen- Estruch, Quan Vuong, Andre He, Vivek Myers, Kuan Fang, Chelsea Finn, and Sergey Levine. Bridgedata v2: A dataset for robot learning at scale. In Conference on Robot Learning ...

    2023

  55. [63]

    The unscented kalman filter for nonlinear estimation

    Eric A Wan and Rudolph Van Der Merwe. The unscented kalman filter for nonlinear estimation. In Proceedings of the IEEE 2000 adaptive systems for signal processing, commu- nications, and control symposium (Cat. No. 00EX373), pages 153–158. Ieee, 2000. 2, 3, 4, 10

    2000

  56. [64]

    Self-supervising action recog- nition by statistical moment and subspace descriptors

    Lei Wang and Piotr Koniusz. Self-supervising action recog- nition by statistical moment and subspace descriptors. In Proceedings of the 29th ACM international conference on multimedia, pages 4324–4333, 2021. 8

    2021

  57. [65]

    Depth any- thing v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth any- thing v2. arXiv:2406.09414, 2024. 5, 10

    2024 arXiv

  58. [66]

    Latent action pretrain- ing from videos

    Seonghyeon Ye, Joel Jang, Byeongguk Jeon, Sejune Joo, Jianwei Yang, Baolin Peng, Ajay Mandlekar, Reuben Tan, Yu-Wei Chao, Bill Yuchen Lin, et al. Latent action pretrain- ing from videos. arXiv preprint arXiv:2410.11758, 2024. 2

    2024 arXiv

  59. [67]

    Metric3d: Towards zero-shot metric 3d prediction from a single image

    Wei Yin, Chi Zhang, Hao Chen, Zhipeng Cai, Gang Yu, Kaixuan Wang, Xiaozhi Chen, and Chunhua Shen. Metric3d: Towards zero-shot metric 3d prediction from a single image. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9043–9053, 2023. 8

    2023

  60. [68]

    Bridg- ing script and animation utilizing a new automatic cine- matography model

    Zixiao Yu, Enhao Guo, Haohong Wang, and Jian Ren. Bridg- ing script and animation utilizing a new automatic cine- matography model. In 2022 IEEE 5th International Con- ference on Multimedia Information Processing and Retrieval (MIPR), pages 268–273. IEEE, 2022. 2

    2022

  61. [69]

    En- abling automatic cinematography with reinforcement learn- ing

    Zixiao Yu, Chenyu Yu, Haohong Wang, and Jian Ren. En- abling automatic cinematography with reinforcement learn- ing. In 2022 IEEE 5th International Conference on Multi- media Information Processing and Retrieval (MIPR) , pages 103–108. IEEE, 2022. 2, 8

    2022

  62. [70]

    Recapture: Gener- ative video camera controls for user-provided videos using masked video fine-tuning

    David Junhao Zhang, Roni Paiss, Shiran Zada, Nikhil Kar- nad, David E Jacobs, Yael Pritch, Inbar Mosseri, Mike Zheng Shou, Neal Wadhwa, and Nataniel Ruiz. Recapture: Gener- ative video camera controls for user-provided videos using masked video fine-tuning. arXiv preprint arXi...

    arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.