Pith. sign in

REVIEW 3 major objections 5 minor 41 references

EgoTouch: On-Body Touch Input Using AR/VR Headset Cameras

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read EgoTouch shows that finger-to-skin touches can be detected from ordinary RGB camera images — the kind already inside AR/VR headsets — by reading the skin deformation and shadow change at the fingertip.

desk verdict Solid, well-evaluated proof-of-concept for RGB-only on-body touch, but the 'works with headset cameras' claim rests on an external camera proxy rather than actual headset streams. read the letter →

arxiv 2509.01786 v1 pith:KQUNHRZE submitted 2025-09-01 cs.HC cs.CVcs.RO

classification cs.HCcs.CVcs.RO
keywords ComputerVisionOn-BodyComputingTouchSurfacesandInteractionAR/VRforceestimationegocentrichandtrackingtransformerbare-hand
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

EgoTouch claims that a standard color (RGB) camera, like the ones already built into AR/VR headsets for passthrough, can reliably tell when a bare finger is touching the user's own skin, purely from the way the skin deforms and shadows shift at the fingertip. The paper reports a 96.4% true-positive touch detection rate with a 5.6% false-positive rate, force estimates with 6.8% mean absolute error, and inference under 0.6 ms on a mobile processor, all with no sensors worn by the user and no per-user calibration. If true, this removes the biggest practical barrier to on-skin interfaces: prior systems needed worn sensors, depth cameras, or fixed lighting. The result matters because on-body touch is faster and more comfortable than in-air gestures, and headsets already carry the required camera hardware.

What carries the argument

The load-bearing object is the localized fingertip patch: roughly 4x4 cm of skin around the fingertip, normalized by hand-pose keypoints to a fixed 100x100 pixel orientation and scale, then classified by a hybrid vision transformer (FastViT T8 backbone, structurally reparameterized to 3.8M parameters). The patch carries the deformation-and-shadow signal that distinguishes touch from near-hover, and the normalization makes that signal comparable across hand sizes, distances, and users. A per-finger state machine converts the frame predictions into touch-down, touch-up, and click events.

What would settle it

Give the trained model live or recorded frames from a current headset's actual RGB passthrough cameras while a user wears the same fingertip ground-truth sensor, and compare true-positive and false-positive rates to the reported 96.4% and 5.6%; a material drop would show the result does not transfer.

Watch

Extended reading notes

Core claim

The central discovery is that finger-to-skin contact leaves a visible signature in ordinary RGB video: the skin deforms locally, changing shading and the finger's shadow, even for subtle touches. EgoTouch chains a 21-keypoint hand tracker, a normalized 100x100 fingertip patch, and a 3.8-million-parameter vision transformer that outputs touch/hover and force. In leave-one-participant-out testing across 15 people, four body sites, indoor and outdoor light from 12 to 35,000 lux, and walking, it reached 96.4% true positives, 5.6% false positives, 6.8% force error, and 95.6% click accuracy. The authors claim this is the first on-body system to combine no user instrumentation, no calibration, and

Load-bearing premise

The accuracy numbers were measured with an external wide-angle camera fastened to a head strap, and the claim that EgoTouch works with cameras already inside real headsets assumes that external camera's image quality and placement match those internal passthrough cameras.

Editorial extensions

If this is right

  • On-skin touch input can run on unmodified commercial headsets using only the RGB passthrough cameras they already contain.
  • Touch interactions can carry continuous force metadata, enabling pressure-sensitive controls similar to 3D Touch.
  • The approach generalizes across skin tones, hair densities, body locations, lighting from 12 to 35,000 lux, and walking motion without calibration.
  • Inference cost (0.51-0.75 ms per frame) is small enough to run as a background process at 90+ FPS on mobile hardware.
  • Multiple active fingers are processed in parallel, so multitouch on the skin is possible.

Reading between the lines

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

  • Because the model's input is a normalized fingertip patch, the same deformation-and-shadow cue likely transfers to other deformable surfaces such as clothing, extending on-body input beyond bare skin.
  • The reported accuracy was measured with an external camera strapped to the head; the headline numbers are an upper bound on transfer until tested on the actual passthrough streams of a consumer headset, which the authors could not access.
  • The paper's own bottleneck analysis suggests that improving hand tracking would raise overall EgoTouch performance more than improving the touch classifier.
  • The visible-signal account predicts that extreme low-light or glare conditions that wash out shading cues will degrade touch detection; the study's 12-lux minimum is a testable lower bound.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. EgoTouch detects finger-to-skin touch and estimates press force from RGB camera images by learning local skin-deformation and shadow cues. The pipeline uses MediaPipe hand keypoints to extract normalized fingertip patches, feeds them into a FastViT-based hybrid vision transformer, and exposes touch/force output through a per-finger state machine. The authors validate with 15 participants under a leave-one-participant-out scheme, spanning indoor/outdoor lighting (12–35,000 lux), Fitzpatrick skin types II–VI, several arm-hair densities, four touch locations, and four touch types, with FSR/capacitive ground truth. Headline results: 96.4% true positive rate, 5.6% false positive rate, 6.8% force MAE, 95.6% same-frame click accuracy, and <0.6 ms inference on an Apple M2. The paper claims this approach works with RGB cameras like those already integrated into modern XR headsets.

Significance. The work addresses a practical gap: prior RGB-only touch-sensing systems either instrument the user, rely on depth cameras, or work only on rigid surfaces. EgoTouch is the first to combine uninstrumented bare-hand input, a calibration-free user-independent model, force estimation, and real-time performance on mobile-class hardware. The evaluation design is a strength: held-out participants, strict same-frame event matching, diverse lighting and skin tones, and transparent reporting of synchronization corrections. If the accuracy transfers to actual headset passthrough cameras, this is a strong contribution to on-body HCI. However, the reported results are based on a proxy camera, and the robustness to walking is not documented in the analysis, which tempers the significance.

major comments (3)
  1. [§4.1, Abstract, §1] The central deployment claim rests on an untested camera-transfer assumption. All study data were collected with an external Insta360 One R mounted on a head strap, not with the RGB passthrough cameras of Quest 3/Vision Pro (which were inaccessible, §3.1). The abstract and intro state the method works with 'cameras like those already integrated into all modern XR headsets.' Actual headset cameras differ in optical distortion, exposure/auto-gain, rolling shutter, frame rate, and placement relative to the user's eyes. The patch extraction (§3.3) uses MediaPipe keypoints but does not undistort or account for camera intrinsics; hence the learned shadow/deformation features may not transfer. Please either (a) evaluate on at least one actual headset passthrough stream, (b) provide a cross-camera generalization experiment (e.g., train on one camera, test on another), or (c) explicitly limit the
  2. [§4.2, §5.2, §5.8] The headline frame-level accuracies include a post hoc global +33.3 ms offset to the ground-truth stream. This offset is a free parameter chosen after data collection, and the fully unaligned accuracy is not reported; §5.8 reports only the fixed-latency accuracy (already offset) and the DTW-corrected accuracy. As a result, the reader cannot separate the model's true temporal performance from the alignment procedure. Please report the unaligned frame-wise accuracy and provide evidence (e.g., hardware latency measurements) that the +33.3 ms offset is a constant of the apparatus rather than a best-fit on the evaluation data. If a range of offsets was tried, state this explicitly.
  3. [Abstract, §5] The claim of robustness to body motion—'e.g., input while walking' in the Abstract—is not supported by any reported analysis. The protocol allowed sitting, standing, and walking (§4.4), but the Bayesian factors in Table 2 do not include a motion condition, and no accuracy breakdown is given for walking vs. stationary frames. Since walking introduces motion blur and global egomotion that could affect the deformation cues, please report a motion-condition breakdown or remove/qualify the walking claim.
minor comments (5)
  1. [§3.1 vs §4.1] The proof-of-concept hardware description in §3.1 names an ELP USBGS1200P01-L21 camera, while §4.1 says the study used an Insta360 One R. Please clarify which camera was used for the user study and why the earlier section mentions a different model.
  2. [§5.8] The sentence 'We believe the latter number is a more faithful representation...' is in tension with the decision to report 94.9% as the main result. The paper would be clearer if the main result and the DTW-corrected result were presented as equally legitimate perspectives, with the unaligned result as a lower bound.
  3. [§6] The Limitations section does not mention the camera-proxy issue; adding it would help readers calibrate the deployment claim.
  4. [§4.4] Participants were asked to 'look at the location they were touching' to ensure both hands are visible. This may not reflect natural AR/VR use, where users often look elsewhere. Please note this as a limitation or justify why it does not affect generalization.
  5. [§4.5] The reference to 'Figure 5' in the cross-reference 'see Section 3.3 and Figure 5' appears to be a typo; Figure 5 shows skin tones and hair density, not the patch extraction pipeline.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: EgoTouch's core accuracy claims are genuine held-out predictions, not reductions to fitted inputs.

full rationale

The paper's central derivation is empirical and self-contained. Touch detection and force estimation models are trained with leave-one-participant-out cross-validation (Section 4.5), so the reported 96.4% true-positive rate, 5.6% false-positive rate, and 6.8% force MAE are measured on participants whose data never entered training. No fitted model parameter is reused as a prediction. The only post hoc alignment, the +33.3 ms global offset and the DTW-corrected numbers in Section 5.8, adjusts evaluation-harness synchronization rather than model output, and the paper explicitly reports both uncorrected and corrected versions, so it does not disguise a fitted quantity as a prediction. The claim that the camera is 'like those already integrated' into XR headsets is an untested domain-transfer assumption (the evaluation used an external Insta360 One R, as disclosed in Sections 3.1 and 4.1), but this is an external-validity concern, not circularity: the accuracy numbers were not constructed from the headset-camera premise. Self-citations to the authors' prior work (OmniTouch, DIRECT, FarOut Touch, LumiWatch, ActiTouch, SkinTrack, etc.) appear as related work and comparison points, not as load-bearing justification for EgoTouch's mechanism or results. There is no uniqueness theorem, no ansatz imported by self-citation, and no renaming of a known result as a new contribution. The derivation chain from RGB patch to touch/force output is supported by an independent ground-truth sensor and out-of-sample evaluation.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

This is an empirical ML systems paper, not a derivation. The main free parameters are the post-hoc alignment choices in the evaluation, not the model weights; the model itself is trained on held-out participants, so out-of-sample evaluation is genuine. No invented entities are introduced.

free parameters (2)
  • Global synchronization offset = +33.3 ms (1 frame at 30 FPS)
    Applied post hoc to the ground truth sensor stream to better align with camera frames in the main evaluation (Section 4.2). This hand-set value shifts all ground truth labels by one frame and affects the reported accuracy.
  • DTW correction window = ±4 frames (±133 ms)
    Used in Section 5.8 to compute the alternative frame-wise accuracy of 97.3% vs. 94.9% raw. Adds per-event time flexibility in scoring that is not present in the deployed system.
assumptions (3)
  • domain assumption MediaPipe Hand keypoints are accurate enough for patch extraction and normalization in ego-centric views.
    Sections 3.2-3.3 use MediaPipe's 21 keypoints to gate and crop input patches; the authors note MediaPipe errors in Section 6, so the result depends on an external tool they do not control.
  • domain assumption Skin deformation from touch produces a detectable, learnable change in shading and shadow under ambient lighting.
    Core physical premise of the approach (Section 1, Figure 2); if the visual cue is inconsistent, the trained model cannot generalize.
  • domain assumption The fingertip-mounted FSR/capacitive ground truth sensor measures touch and force without biasing the camera-visible appearance.
    Section 4.1; the sensor is taped under the finger and claimed invisible, but tape and sensor could alter deformation and shading in a way the model learns.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EgoTouch: On-Body Touch Input Using AR/VR Headset Cameras." pith.science (2026). https://pith.science/paper/KQUNHRZE

@misc{pith2026250901786,
  author       = {Pith},
  title        = {Pith review of: EgoTouch: On-Body Touch Input Using AR/VR Headset Cameras},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KQUNHRZE}},
  note         = {Machine review of arXiv:2509.01786}
}
read the original abstract

In augmented and virtual reality (AR/VR) experiences, a user's arms and hands can provide a convenient and tactile surface for touch input. Prior work has shown on-body input to have significant speed, accuracy, and ergonomic benefits over in-air interfaces, which are common today. In this work, we demonstrate high accuracy, bare hands (i.e., no special instrumentation of the user) skin input using just an RGB camera, like those already integrated into all modern XR headsets. Our results show this approach can be accurate, and robust across diverse lighting conditions, skin tones, and body motion (e.g., input while walking). Finally, our pipeline also provides rich input metadata including touch force, finger identification, angle of attack, and rotation. We believe these are the requisite technical ingredients to more fully unlock on-skin interfaces that have been well motivated in the HCI literature but have lacked robust and practical methods.

Figures

Figures reproduced from arXiv: 2509.01786 by the authors.

Figure 1
Figure 1. Using RGB cameras already present in modern XR headsets for passthrough, EgoTouch can enable on-body touch [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. A finger hovering just above the skin (A) vs. touch [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Software pipeline overview. First, hand keypoints are extracted from the RGB camera view (A; shown in blue and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Our ground truth sensor (close-up in upper-right) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Example skin tones and hair densities collected in our user study. Scales based on [6, 32]. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Frame-wise touch classification accuracy vs. skin [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Frame-wise touch classification accuracy vs. touch [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 9
Figure 9. Figure 9: Frame-wise touch classification accuracy (A) and [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 8
Figure 8. Figure 8: Our study included diverse lighting conditions, in [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 10
Figure 10. Figure 10: A real-world, ten-second sequence from our user study containing eight long press events of varying durations. [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

41 extracted references · 21 canonical work pages

  1. [1]

    Android Developers. 2024. Input Events API. https://developer.android.com/ develop/ui/views/touch-and-input/input-events

  2. [2]

    Joanna Bergström and Kasper Hornbæk. 2019. Human–Computer Interaction on the Skin. ACM Comput. Surv. 52, 4, Article 77 (aug 2019), 14 pages. https: //doi.org/10.1145/3332166

  3. [3]

    Brewster and Michael Hughes

    Stephen A. Brewster and Michael Hughes. 2009. Pressure-based text entry for mobile devices. In Proceedings of the 11th International Conference on Human- Computer Interaction with Mobile Devices and Services (Bonn, Germany) (Mobile- HCI ’09). Association for Computing Machinery, New York, NY, USA, Article 9, 4 pages. https://doi.org/10.1145/1613858.1613870

  4. [4]

    Jesse Burstyn, Paul Strohmeier, and Roel Vertegaal. 2015. DisplaySkin: Exploring Pose-Aware Displays on a Flexible Electrophoretic Wristband. In Proceedings of the Ninth International Conference on Tangible, Embedded, and Embodied Interac- tion (Stanford, California, USA) (TEI ’15). Association for Computing Machinery, New York, NY, USA, 165–172. https://...

  5. [5]

    Camille Dupré, Caroline Appert, Stéphanie Rey, Houssem Saidi, and Emmanuel Pietriga. 2024. TriPad: Touch Input in AR on Ordinary Surfaces with Hand Tracking Only. In CHI 2024 - The 42nd SIGCHI conference on Human Factors in computing systems, ACM (Ed.). ACM, Honolulu, HI, USA, United States. https: //doi.org/10.1145/3613904.3642323

  6. [6]

    Thomas B Fitzpatrick. 1975. Soleil et peau. Journal de Medecine Esthetique 2 (1975), 33–34

  7. [7]

    Collins, Chengcheng Tang, Christopher D

    Patrick Grady, Jeremy A. Collins, Chengcheng Tang, Christopher D. Twigg, Kunal Aneja, James Hays, and Charles C. Kemp. 2024. PressureVision++: Estimating Fingertip Pressure from Diverse RGB Images. (2024), 8683–8693. https://doi.org/ 10.1109/WACV57701.2024.00850

  8. [8]

    Twigg, Chengde Wan, James Hays, and Charles C

    Patrick Grady, Chengcheng Tang, Samarth Brahmbhatt, Christopher D. Twigg, Chengde Wan, James Hays, and Charles C. Kemp. 2022. PressureVision: Estimat- ing Hand Pressure from a Single RGB Image. In Computer Vision - ECCV 2022 - 17th European Conference, Tel A viv, Israel, October 23-27, 2022, Proceedings, Part VI (Lecture Notes in Computer Science, Vol. 13...

Show all 41 references
  1. [10]

    Sean Gustafson, Bernhard Rabe, and Patrick Baudisch. 2013. Understanding palm- based imaginary interfaces: the role of visual and tactile cues when browsing. In 2013 ACM SIGCHI Conference on Human Factors in Computing Systems, CHI ’13, Paris, France, April 27 - May 2, 2013 , W...

  2. [12]

    Chris Harrison, Desney Tan, and Dan Morris. 2010. Skinput: appropriating the body as an input surface. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Atlanta, Georgia, USA) (CHI ’10). Association for Computing Machinery, New York, NY, USA, 453–4...

  3. [13]

    Johansson and J Randall Flanagan

    Roland S. Johansson and J Randall Flanagan. 2009. Coding and use of tactile sig- nals from the fingertips in object manipulation tasks.Nature Reviews Neuroscience 10 (2009), 345–359. https://api.semanticscholar.org/CorpusID:17298704

  4. [14]

    Hsin-Liu (Cindy) Kao, Christian Holz, Asta Roseway, Andres Calvo, and Chris Schmandt. 2016. DuoSkin: rapidly prototyping on-skin user interfaces using skin-friendly materials. In Proceedings of the 2016 ACM International Symposium on Wearable Computers (Heidelberg, Germany) (I...

  5. [15]

    Wolf Kienzle, Eric Whitmire, Chris Rittaler, and Hrvoje Benko. 2021. ElectroRing: Subtle Pinch and Touch Detection with a Ring. (2021), 3:1–3:12. https://doi.org/ 10.1145/3411764.3445094

  6. [16]

    Taejun Kim, Amy Karlson, Aakar Gupta, Tovi Grossman, Jason Wu, Parastoo Abtahi, Christopher Collins, Michael Glueck, and Hemant Bhaskar Surale. 2023. STAR: Smartphone-analogous Typing in Augmented Reality. InProceedings of the 36th Annual ACM Symposium on User Interface Softwa...

  7. [17]

    Yuki Kubo, Yuto Koguchi, Buntarou Shizuki, Shin Takahashi, and Otmar Hilliges

  8. [18]

    Hudson, and Chris Harrison

    Gierad Laput, Robert Xiao, Xiang ’Anthony’ Chen, Scott E. Hudson, and Chris Harrison. 2014. Skin buttons: cheap, small, low-powered and clickable fixed- icon laser projectors. In Proceedings of the 27th Annual ACM Symposium on User Interface Software and Technology(Honolulu, H...

  9. [20]

    Shu-Yang Lin, Chao-Huai Su, Kai-Yin Cheng, Rong-Hao Liang, Tzu-Hao Kuo, and Bing-Yu Chen. 2011. Pub - point upon body: exploring eyes-free interaction and methods on an arm. In Proceedings of the 24th Annual ACM Symposium on User Interface Software and Technology, Santa Barbar...

  10. [21]

    Tan, Dan Morris, Shwetak N

    Adiyan Mujibiya, Xiang Cao, Desney S. Tan, Dan Morris, Shwetak N. Patel, and Jun Rekimoto. 2013. The sound of touch: on-body touch and gesture sensing based on transdermal ultrasound propagation (ITS ’13). Association for Computing Ma- chinery, New York, NY, USA, 189–198. http...

  11. [23]

    Ryosuke Ono, Shunsuke Yoshimoto, and Kosuke Sato. 2013. Palm+Act: operation by visually captured 3D force on palm. InSIGGRAPH Asia 2013 Emerging Technolo- gies (Hong Kong, Hong Kong) (SA ’13). Association for Computing Machinery, New York, NY, USA, Article 14, 3 pages. https:/...

  12. [24]

    D. T. V. Pawluk and R. D. Howe. 1999. Dynamic Contact of the Human Fingerpad Against a Flat Surface. Journal of Biomechanical Engineering 121, 6 (12 1999), 605–611. https://doi.org/10.1115/1.2800860

  13. [25]

    Ivan Poupyrev, Chris Harrison, and Munehiko Sato. 2012. Touché: touch and gesture sensing for the real world. In Proceedings of the 2012 ACM Conference on Ubiquitous Computing (Pittsburgh, Pennsylvania) (UbiComp ’12). Association for Computing Machinery, New York, NY, USA, 536...

  14. [26]

    Vivian Shen, James Spann, and Chris Harrison. 2021. FarOut Touch: Extending the Range of ad hoc Touch Sensing with Depth Cameras. In Proceedings of the 2021 ACM Symposium on Spatial User Interaction (Virtual Event, USA) (SUI ’21). Association for Computing Machinery, New York,...

  15. [27]

    Yilei Shi, Haimo Zhang, Kaixing Zhao, Jiashuo Cao, Mengmeng Sun, and Suranga Nanayakkara. 2020. Ready, Steady, Touch!: Sensing Physical Contact with a Finger-Mounted IMU. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 4, 2 (2020), 59:1–59:25. https://doi.org/10.1145/3397309

  16. [28]

    Srinath Sridhar, Anders Markussen, Antti Oulasvirta, Christian Theobalt, and Sebastian Boring. 2017. WatchSense: On- and Above-Skin Input Sensing through a Wearable Depth Sensor. InProceedings of the 2017 CHI Conference on Human Factors in Computing Systems (Denver, Colorado, ...

  17. [29]

    Froehlich

    Lee Stearns, Uran Oh, Leah Findlater, and Jon E. Froehlich. 2017. TouchCam: Realtime Recognition of Location-Specific On-Body Gestures to Support Users with Visual Impairments. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 1, 4 (2017), 164:1–164:23. https://doi.org/10....

  18. [30]

    Paul Streli, Jiaxi Jiang, Juliete Rossie, and Christian Holz. 2023. Structured Light Speckle: Joint Ego-Centric Depth Estimation and Low-Latency Contact Detection via Remote Vibrometry. InProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, U...

  19. [31]

    Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, Oncel Tuzel, and Anurag Ranjan. 2023. FastViT: A Fast Hybrid Vision Transformer using Structural Reparameterization. In IEEE/CVF International Conference on Computer Vision, ICCV 2023, Paris, France, October 1-6, 2023 . IEEE...

  20. [32]

    LA von Schuckmann, MC Hughes, Adèle C Green, and JC van der Pols. 2016. Fore- arm hair density and risk of keratinocyte cancers in Australian adults. Archives of dermatological research 308 (2016), 617–624

  21. [33]

    Anandghan Waghmare, Youssef Ben Taleb, Ishan Chatterjee, Arjun Narendra, and Shwetak N. Patel. 2023. Z-Ring: Single-Point Bio-Impedance Sensing for Gesture, EgoTouch: On-Body Touch Input Using AR/VR Headset Cameras UIST ’24, October 13–16, 2024, Pittsburgh, PA, USA Touch, Obje...

  22. [34]

    Cheng-Yao Wang, Wei-Chen Chu, Po-Tsung Chiu, Min-Chieh Hsiu, Yih-Harn Chiang, and Mike Y. Chen. 2015. PalmType: Using Palms as Keyboards for Smart Glasses. In Proceedings of the 17th International Conference on Human-Computer Interaction with Mobile Devices and Services, Mobil...

  23. [35]

    Martin Weigel, Tong Lu, Gilles Bailly, Antti Oulasvirta, Carmel Majidi, and Jürgen Steimle. 2015. iSkin: Flexible, Stretchable and Visually Customizable On-Body Touch Sensors for Mobile Computing. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing S...

  24. [36]

    Martin Weigel, Vikram Mehta, and Jürgen Steimle. 2014. More than touch: understanding how people use skin as an input surface for mobile computing. In CHI Conference on Human Factors in Computing Systems, CHI’14, Toronto, ON, Canada - April 26 - May 01, 2014 , Matt Jones, Phil...

  25. [37]

    Andrew D. Wilson. 2005. PlayAnywhere: a compact interactive tabletop projection-vision system. In Proceedings of the 18th Annual ACM Symposium on User Interface Software and Technology (Seattle, WA, USA) (UIST ’05). Associa- tion for Computing Machinery, New York, NY, USA, 83–...

  26. [38]

    Robert Xiao, Teng Cao, Ning Guo, Jun Zhuo, Yang Zhang, and Chris Harrison

  27. [39]

    Robert Xiao, Scott Hudson, and Chris Harrison. 2016. DIRECT: Making Touch Tracking on Ordinary Surfaces Practical with Hybrid Depth-Infrared Sensing. In Proceedings of the 2016 ACM International Conference on Interactive Surfaces and Spaces (Niagara Falls, Ontario, Canada) (IS...

  28. [40]

    Robert Xiao, Julia Schwarz, and Chris Harrison. 2015. Estimating 3D Finger Angle on Commodity Touchscreens. In Proceedings of the 2015 International Conference on Interactive Tabletops & Surfaces, ITS 2015, Funchal, Portugal, November 15- 18, 2015, Nuno Jardim Nunes, Enrico Co...

  29. [41]

    Fan Zhang, Valentin Bazarevsky, Andrey Vakunov, Andrei Tkachenka, George Sung, Chuo-Ling Chang, and Matthias Grundmann. 2020. MediaPipe Hands: On-device Real-time Hand Tracking. arXiv:2006.10214

  30. [42]

    Ng, Hrvoje Benko, and Chris Harrison

    Yang Zhang, Wolf Kienzle, Yanjun Ma, Shiu S. Ng, Hrvoje Benko, and Chris Harrison. 2019. ActiTouch: Robust Touch Detection for On-Skin AR/VR Interfaces. In Proceedings of the 32nd Annual ACM Symposium on User Interface Software and Technology (New Orleans, LA, USA) (UIST ’19)....

  31. [43]

    Yang Zhang, Junhan Zhou, Gierad Laput, and Chris Harrison. 2016. SkinTrack: Using the Body as an Electrical Waveguide for Continuous Finger Tracking on the Skin. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (San Jose, California, USA) (CHI ’1...

  32. [2018]

    In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, CHI 2018, Montreal, QC, Canada, April 21-26, 2018, Regan L

    LumiWatch: On-Arm Projected Graphics and Touch Input. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, CHI 2018, Montreal, QC, Canada, April 21-26, 2018, Regan L. Mandryk, Mark Hancock, Mark Perry, and Anna L. Cox (Eds.). ACM, 95. https://doi.or...

  33. [2019]

    In Proceedings of the 21st International Conference on Human- Computer Interaction with Mobile Devices and Services (Taipei, Taiwan)(MobileHCI ’19)

    AudioTouch: Minimally Invasive Sensing of Micro-Gestures via Active Bio- Acoustic Sensing. In Proceedings of the 21st International Conference on Human- Computer Interaction with Mobile Devices and Services (Taipei, Taiwan)(MobileHCI ’19). Association for Computing Machinery, ...

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.