Pith. sign in

REVIEW 5 major objections 6 minor 45 references

VolE: A Point-cloud Framework for Food 3D Reconstruction and Volume Estimation

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read VolE claims that a reference-free mobile phone scan can estimate food volume with a mean absolute percentage error of 2.22% on the MTF benchmark, by using AR-tracked camera positions to give the 3D reconstruction real-world scale.

desk verdict Useful integration paper with a load-bearing description bug: the all-views intersection masking rule contradicts the 360° capture protocol, the headline MAPE disagrees with Table III, and the AR-to-COLMAP scale alignment is underspecified. read the letter →

arxiv 2505.10205 v1 pith:UBOL7LCE submitted 2025-05-15 cs.CV

Umair Haroon , Ahmad AlMughrabi , Thanasis Zoumpekas , Ricardo Marques , Petia Radeva This is my paper · ORCID
classification cs.CV
keywords foodvolumeestimation3Dreconstructionpointcloudaugmentedrealitystructurefrommotionvideosegmentationdietaryassessmentmobilephonescanning
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

VolE aims to establish that an ordinary AR-capable smartphone, moved freely around a plate of food, can estimate food volume precisely without a reference object, a depth sensor, or training on food shapes. The paper's load-bearing idea is that the phone's augmented-reality pose tracking gives each captured frame a 3D camera location in real-world units, and structure-from-motion refinement keeps that scale when the frames are reconstructed into a point cloud and mesh. On the MTF benchmark the paper claims a 2.22% mean absolute percentage error, and on its new 21-item Foodkit dataset it reports a 1.22% mean absolute error. The framework also outputs a refined 3D mesh, so testable claims include both volume accuracy and geometric fidelity, measured by Chamfer distance. If the scale assumption holds, this would remove the main practical barrier to everyday dietary volume logging: you just film the food.

What carries the argument

The load-bearing mathematics is the closed-mesh volume identity from the divergence theorem, $V = \frac{1}{6}\sum_{k=1}^{N} v_k^1 \cdot (v_k^2 \times v_k^3)$, which sums the signed volumes of tetrahedra formed by each triangular face and the origin. What makes that identity usable in real-world units is the metric scale supplied by the phone's AR pose tracker at capture time and preserved through structure-from-motion refinement; every downstream object—the dense point cloud, the masked food point set, and the final mesh—inherits that scale, so no reference object or depth sensor is needed. The mask-generation stage supplies per-frame food regions, and point-cloud masking retains only 3D points whose projection lies inside the mask in every view, so the tetrahedron sum runs over food geometry rather than background.

What would settle it

Record a rigid object of known volume, such as a 200 ml measuring cup, with VolE under slow steady motion and under fast jerky motion, then compare estimated volumes to the true volume; if the fast capture or a long camera path pushes the error well above the reported 2-3%, the AR scale and drift assumption fails.

Watch

Extended reading notes

Core claim

The central discovery, on the paper's own terms, is that mobile augmented-reality camera poses can serve as the metric scaffold that replaces reference objects and depth information in food volume estimation. The pipeline captures a video of the food while the phone logs a 3D pose per frame, refines those poses with structure from motion, generates per-frame food masks with a video segmentation model, projects and intersects those masks into the scene point cloud to keep only food points, builds and refines a closed mesh, and computes volume as the sum of signed tetrahedra from the divergence theorem. The paper's evidence is the reported error rates and Chamfer distances across three datasets, including lower mean absolute percentage error than the compared baselines on the MTF benchmark and a lower mean Chamfer distance than several established reconstruction methods on the DTU multi-view benchmark.

Load-bearing premise

The whole method assumes the phone's augmented-reality tracker reports real-world distances accurately enough that no reference object is needed, and that these distances do not drift while the user moves the camera around the food.

Editorial extensions

If this is right

  • A user can log food volume on an ordinary phone by filming the meal, since the AR tracker supplies scale and the pipeline adds masks automatically.
  • The same captured video yields both a volume estimate and a refined 3D mesh, so dietary logging can be combined with shape-level analysis of the meal.
  • Selecting frames by visual dissimilarity can cut reported processing time by more than 90% while keeping accuracy above 97%, which points toward near-real-time use.
  • Low-textured foods and unbounded scenes, which the paper identifies as hard cases for NeRF-style reconstruction, are the cases the framework is designed to handle.
  • The new 21-object Foodkit dataset, once released, gives the field food items with water-displacement ground-truth volumes and masses for benchmarking future reference-free methods.

Reading between the lines

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

  • Because nothing in the pipeline is food-specific, the same reference-free AR-scaffolded reconstruction could be transferred to other free-motion volume tasks, such as parcel sizing or wound measurement; the paper does not claim this.
  • A testable consequence of the AR-scale premise is that volume accuracy should degrade with path length or fast motion as AR tracking drifts; the paper reports neither a drift analysis nor an alignment check against a known object size.
  • Substituting manually drawn masks for the video segmentation output would isolate how much of the reported error comes from segmentation versus geometry, an ablation the paper does not run.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper presents VolE, a mobile-device-based framework for food volume estimation. The pipeline captures images and camera poses with ARKit/ARCore, refines poses with COLMAP, generates food masks with FoodMem, extracts a segmented point cloud, reconstructs and refines a mesh with OpenMVS, and computes volume via the divergence theorem. The authors introduce a new Foodkit dataset of 21 food objects with water-displacement ground truth and report experiments on Foodkit, MTF, and DTU. The headline claim is that VolE achieves 2.22% MAPE on MTF and 1.22% mean absolute error on Foodkit, outperforming prior volume-estimation methods without requiring reference objects or depth sensors.

Significance. If the technical issues are resolved, the idea of using AR-based pose priors to obtain metric scale without reference objects is practically valuable for mobile dietary assessment, and the Foodkit dataset addresses a real gap in food volume benchmarks. The paper is well-structured and includes useful robustness checks, such as repeated runs in the ablation tables. However, the central comparative claim is currently weakened by several load-bearing inconsistencies: the point-cloud masking rule as written is incompatible with 360-degree capture, the headline MAPE does not match the reported table, and the reference-free scaling procedure is not specified. These issues need to be addressed before the results can be considered reproducible.

major comments (5)
  1. [Section III-E (Point Cloud Masking)] The masking rule is defined as P = ∩_{j=1}^{N_I} M_j, meaning a scene point is retained only if its projection lies inside the food mask in every frame. For a 360-degree free-motion capture (Sec. IV-C-1) and for the multi-view MTF scenes, most surface points are occluded or out of view in a large fraction of frames, so this intersection will be empty or far too sparse to support the Delaunay/graph-cut/marching-cubes mesh pipeline in Sec. III-F. The volume errors in Tables II and III therefore cannot follow from the stated algorithm. Please replace the intersection rule with a visibility-aware rule (e.g., retain a point if it projects inside the mask in all frames where the point is actually visible, or in a sufficient subset) and re-run or re-report the affected experiments.
  2. [Abstract and Section IV-D.2 vs. Table III] The abstract and Sec. IV-D.2 state that VolE achieves 2.22% MAPE on the MTF dataset, but Table III reports a MAPE of 3.08%, and averaging the 13 listed per-scene error percentages for VolE gives 3.08%. This is a direct numerical inconsistency in the headline claim. Please correct the abstract/text or explain which subset of scenes the 2.22% figure refers to.
  3. [Sections III-B, IV-C-1, and IV-C-2] The reference-free scale claim is not supported as written. Section III-B states that ARKit/ARCore positions provide metric scale, and Sec. IV-C-1 says ARKit coordinates are used to scale the reconstruction, but no alignment procedure between the AR poses and the COLMAP SfM reconstruction is specified (e.g., similarity-transform estimation, which frames are used, and how AR drift is handled). Furthermore, Sec. IV-C-2 states that on MTF 'we used the reference board to scale the reconstructed scenes to their original size,' which, if applied to VolE, would contradict the reference-free claim. Please specify the scale calibration for both datasets and clarify whether the MTF results are reference-free.
  4. [Section IV-C-1 and Table II] The Foodkit ground-truth volumes were measured by water displacement with a stated error margin of ±5 mL. Table II reports per-item mean absolute errors as low as 0.30–0.40 mL (e.g., Durum 0.40 mL, Yellow Cane 0.30 mL), far below the precision of the ground-truth reference. These errors should be interpreted against the ±5 mL GT uncertainty, and the paper should avoid implying sub-mL accuracy for volumes whose reference is only known to within ±5 mL.
  5. [Section IV-D and Tables II–IV] The claim of outperforming existing volume estimators 'across multiple datasets' is only demonstrated for volume on MTF (Table III). Foodkit (Table II) contains no baseline volume comparisons, and DTU (Table IV) reports only Chamfer distance, not volume. Please state how the baseline numbers in Table III were obtained (official code, reimplementation, same masks and scaling) and add at least one volume-estimation baseline on Foodkit to support the comparative claim.
minor comments (6)
  1. [Section I (Introduction)] The phrase 'mononuclear data' should be 'monocular data'.
  2. [Section IV-D.2] Two consecutive paragraphs both begin 'Table III presents...' and refer to different content (volume error and Chamfer distance) in the same table; please renumber or rename the table or headings for clarity.
  3. [References] Reference [13] is corrupted: the citation string contains raw BibTeX fragments and a malformed author list. Please fix this entry.
  4. [Table III] The ID column in Table III jumps from 11 to 13 with no row 12; please renumber the scenes consistently.
  5. [Section III-E] The symbol P is used for both the complete scene point cloud and the final segmented object point cloud; please use distinct symbols to avoid confusion.
  6. [Section IV-C-1] The sentence 'the error margin of ±5 mL comes from the scaler' presumably refers to the measuring cylinder or scale used in water displacement; please clarify the wording.

Circularity Check

3 steps flagged · score 4.0 of 10

VolE's volume numbers are anchored to external water-displacement and DTU ground truths, but the written pipeline contains a self-definitional mask-intersection step, a threshold-selected ablation reported as accuracy, and load-bearing self-citations (FoodMem, MVS-Boost).

  1. self definitional [Sec. III-E (Point Cloud Masking), with the Foodkit capture protocol in Sec. IV-C-1]
    "The final segmented object point cloud P is the intersection of all these valid points across all images: P = ∩_{j=1}^{N_I} M_j. This ensures that only points visible and masked in all views are retained, leading to the final segmented point cloud of the required object. ... The data collection process involved capturing each food object in a 360-degree view using an ARKit-based mobile application."

    By defining the segmented object cloud as the intersection over all views, a point survives only if its projection lies inside mask S_j for every frame j. In the 360-degree free-motion captures described for Foodkit, every surface point is out of view or occluded in at least some frames, so its projection is absent from those masks and the point is discarded by construction. The Delaunay/graph-cut/marching-cubes mesh of Sec. III-F and the divergence-theorem volume of Sec. III-H are therefore computed from a cloud that is, by definition, missing the back-side surface. The reported 360-degree reconstructions and the 1.22% / 3.08% / 2.22% MAPE numbers cannot follow from the stated intersection rule; a visibility-aware masking rule would be required but is not specified.

  2. fitted input called prediction [Sec. IV-E-2 (Hamming Distance Experiments)]
    "We systematically tested Hamming distance thresholds from 12 to 1, decreasing by one at each step. ... The results show that a Hamming distance threshold of 2 achieves the best balance between accuracy and efficiency, using 486 images and yielding a mean absolute error of 0.07 ml with 99.93% accuracy."

    The threshold is selected after enumerating all values on the same single apple scene (ground truth 175 ml) that is then used to report the accuracy. There is no held-out scene, no cross-validation, and no separate test set for the selection. Reporting the minimum of the threshold-error curve as '99.93% accuracy' is reporting a hyperparameter fitted by the evaluation metric as if it were an independent prediction; the 0.07 ml error is forced by the selection procedure rather than being an out-of-sample estimate.

1 more flagged steps
  1. self citation load bearing [Sec. III-C and Sec. III-F, with Refs [6] and [40]]
    "Simultaneously, FoodMem isolates the object of interest, generating segmented masks S. ... Mesh reconstruction [39] involves creating a mesh from the segmented point cloud P using a Multi-View Stereo (MVS) [40] approach."

    The two components that convert raw RGB frames into the object geometry—the FoodMem video-segmentation masks and the MVS-Boost reconstruction—are arXiv preprints by the present authors. This paper provides no code reproduction, machine-checked verification, or independent evaluation of either component, yet their correctness is load-bearing for every point cloud, mesh, and volume estimate in the experiments. The final ground-truth volumes are external, so this is not a full definitional circle, but the accuracy claim rests in part on an unverified self-citation chain rather than on evidence internal to the paper.

full rationale

VolE does not fit any volume parameter to the ground-truth volumes; the final MAPE values are obtained by comparing divergence-theorem volumes of reconstructed meshes against external water-displacement (Foodkit/MTF) and DTU ground truths, so the core comparison is not definitionally circular. However, three load-bearing steps weaken the derivation chain. First, Sec. III-E defines the segmented point cloud as the intersection of per-view mask projections; combined with the 360-degree captures described in Sec. IV-C-1, this discards every point not visible in all frames, so the mesh and volume are, by construction, computed from an incomplete cloud. Second, the Hamming-distance ablation chooses the threshold that minimizes error on the same apple scene it then reports as 99.93% accuracy, making that number a selected fit rather than an independent prediction. Third, the components that turn raw images into geometry—FoodMem masks and MVS-Boost reconstruction—are arXiv preprints by the same authors and are not reproduced or independently verified in the paper; they are load-bearing for the reported accuracy. These are genuine weaknesses, but they do not make the main result equivalent to its inputs, because the volume targets are external benchmarks and no volume parameter is fitted to them. A score of 4 reflects the partial self-citation chain and the by-construction selection in the ablation, while acknowledging that the central external evaluation retains independent content.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The volume result depends on hand-chosen OpenMVS and frame-selection settings, on the unstated AR-to-SfM scaling procedure, and on the reliability of the authors' own FoodMem segmenter. No new physical entities are introduced. The divergence theorem is the only standard-math input; all other axioms are domain assumptions that need empirical validation.

free parameters (7)
  • OpenMVS max-resolution (Foodkit/MTF) = 512
    Chosen to balance detail and runtime for point cloud masking and meshing.
  • OpenMVS close-holes = 50
    Chosen to close holes in the reconstructed surface; directly affects volume.
  • OpenMVS smooth = 5
    Surface regularization strength; affects geometry and volume.
  • OpenMVS sub-resolution-levels (DTU) = 8
    Set for high-resolution DTU processing.
  • Graph-cut iters (DTU) = 3
    Number of depth estimation iterations for DTU.
  • Geometric iters (DTU) = 2
    Number of geometric consistency iterations for DTU.
  • Hamming distance threshold = 2
    Selected in the ablation as the best accuracy/time trade-off for frame selection on the apple scene; the paper does not state whether this selection is used in the final MTF/Foodkit results.
assumptions (6)
  • domain assumption ARKit/ARCore provide camera poses with metric scale that remain accurate enough during free motion
    Invoked in Sec. III-B and IV-C-1; the paper gives no drift or calibration analysis for free-motion capture.
  • domain assumption The SfM reconstruction can be aligned to the AR poses via a similarity transformation that survives bundle adjustment
    The scaling step in Sec. IV-C-1 assumes AR coordinates can be used to scale the COLMAP output, but the alignment procedure is not specified.
  • domain assumption FoodMem segmentation masks correctly isolate the food in every frame
    Masks drive point cloud masking in Sec. III-E; errors shrink or enlarge the retained point cloud and change the volume.
  • domain assumption The refined mesh is closed and watertight, so the divergence theorem volume formula applies
    Sec. III-H uses the signed tetrahedron formula, which is exact only for closed surfaces; the close-holes and smoothing steps alter the surface.
  • domain assumption COLMAP and OpenMVS produce a geometrically faithful reconstruction for low-texture and specular food surfaces
    The DTU benchmark is not food, and MTF may have texture; food items such as chocolate and lemon rind may violate typical SfM assumptions.
  • standard math Divergence theorem volume formula V = (1/6) * sum_k v1 dot (v2 cross v3) is exact for a closed triangular mesh
    Sec. III-H; standard calculus result, not a weakness.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VolE: A Point-cloud Framework for Food 3D Reconstruction and Volume Estimation." pith.science (2026). https://pith.science/paper/UBOL7LCE

@misc{pith2026250510205,
  author       = {Pith},
  title        = {Pith review of: VolE: A Point-cloud Framework for Food 3D Reconstruction and Volume Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UBOL7LCE}},
  note         = {Machine review of arXiv:2505.10205}
}
read the original abstract

Accurate food volume estimation is crucial for medical nutrition management and health monitoring applications, but current food volume estimation methods are often limited by mononuclear data, leveraging single-purpose hardware such as 3D scanners, gathering sensor-oriented information such as depth information, or relying on camera calibration using a reference object. In this paper, we present VolE, a novel framework that leverages mobile device-driven 3D reconstruction to estimate food volume. VolE captures images and camera locations in free motion to generate precise 3D models, thanks to AR-capable mobile devices. To achieve real-world measurement, VolE is a reference- and depth-free framework that leverages food video segmentation for food mask generation. We also introduce a new food dataset encompassing the challenging scenarios absent in the previous benchmarks. Our experiments demonstrate that VolE outperforms the existing volume estimation techniques across multiple datasets by achieving 2.22 % MAPE, highlighting its superior performance in food volume estimation.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 34 canonical work pages

  1. [1]

    V oleta: One- and few-shot food volume estimation,

    A. AlMughrabi, U. Haroon, R. Marques, and P. Radeva, “V oleta: One- and few-shot food volume estimation,”arXiv preprint arXiv:2407.01717, 2024

    arXiv 2024

  2. [2]

    Guidance for monitoring healthy diets globally,

    UNICEF et al., “Guidance for monitoring healthy diets globally,” online, 2024

    work page 2024

  3. [3]

    A systematic review of dietary data collection methodologies for diet diversity indicators,

    S. Mahal, C. Kucha, E. M. Kwofie, and M. Ngadi, “A systematic review of dietary data collection methodologies for diet diversity indicators,” Frontiers in Nutrition, vol. 11, p. 1195799, 2024

    work page 2024

  4. [4]

    Experience sampling as a dietary assess- ment method: a scoping review towards implementation,

    J. Verbeke and C. Matthys, “Experience sampling as a dietary assess- ment method: a scoping review towards implementation,” International Journal of Behavioral Nutrition and Physical Activity , vol. 21, no. 1, p. 94, 2024

    work page 2024

  5. [5]

    Food recognition and nutritional apps,

    L. A. Rahman, I. Papathanail, L. Brigato, E. K. Spanakis, and S. Mougiakakou, “Food recognition and nutritional apps,” in Diabetes Digital Health, Telehealth, and Artificial Intelligence . Elsevier, 2024, pp. 73–83

    work page 2024

  6. [6]

    FoodMem: Near Real-time and Precise Food Video Segmentation

    A. AlMughrabi, A. Gal ´an, R. Marques, and P. Radeva, “Foodmem: Near real-time and precise food video segmentation,” arXiv preprint arXiv:2407.12121, 2024

    work page Pith review arXiv 2024

  7. [7]

    Bayesian deep learning for semantic segmentation of food images,

    E. Aguilar, B. Nagarajan, B. Remeseiro, and P. Radeva, “Bayesian deep learning for semantic segmentation of food images,” Computers and Electrical Engineering, vol. 103, p. 108380, 2022

    work page 2022

  8. [8]

    Lofi: Long-tailed fine-grained network for food recognition,

    J. M. Rodr ´ıguez-De-Vera, I. G. Estepa, M. Bola ˜nos, B. Nagarajan, and P. Radeva, “Lofi: Long-tailed fine-grained network for food recognition,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 3750–3760

    work page 2024

Show all 45 references
  1. [9]

    Deep ensemble- based hard sample mining for food recognition,

    B. Nagarajan, M. Bola ˜nos, E. Aguilar, and P. Radeva, “Deep ensemble- based hard sample mining for food recognition,” Journal of Visual Communication and Image Representation , vol. 95, p. 103905, 2023

    2023

  2. [10]

    Mobile computer vision-based applications for food recognition and volume and calorific estimation: A systematic review,

    L. M. Amugongo, A. Kriebitz, A. Boch, and C. L ¨utge, “Mobile computer vision-based applications for food recognition and volume and calorific estimation: A systematic review,” in Healthcare, vol. 11, no. 1. MDPI, 2022, p. 59

    2022

  3. [11]

    An end-to-end food image analysis system,

    J. He, R. Mao, Z. Shao, J. L. Wright, D. A. Kerr, C. J. Boushey, and F. Zhu, “An end-to-end food image analysis system,” arXiv preprint arXiv:2102.00645, 2021

    2021 arXiv

  4. [12]

    An integrated system for mobile image-based dietary assessment,

    Z. Shao, Y . Han, J. He, R. Mao, J. Wright, D. Kerr, C. J. Boushey, and F. Zhu, “An integrated system for mobile image-based dietary assessment,” in Proceedings of the 3rd Workshop on AIxFood , 2021, pp. 19–23

    2021

  5. [13]

    Metafood cvpr 2024 challenge on physically informed 3d food reconstruction: Methods and results,

    Y . He, Jiangpeng an@inproceedingshou2024low, title=Low Texture 3D Reconstruction Based on Deep Learning Hierarchical Networks, au- thor=Hou, Yaohui and Yang, Zhuo and Hu, Yang and Wang, Lijun, booktitle=2024 2nd International Conference on Algorithm, Image Pro- cessing and Ma...

    2024 arXiv

  6. [14]

    Stan- dardised and objective dietary intake assessment tool (sodiat): Protocol of a dual-site dietary intervention study to integrate dietary assessment methods,

    E. Bobokhidze, M. Weech, K. Petropoulou, T. Wilson, J. Pugh, R. Fal- laize, I. Garcia-Perez, F. P.-W. Lo, A. R Solis, J. Vickar et al. , “Stan- dardised and objective dietary intake assessment tool (sodiat): Protocol of a dual-site dietary intervention study to integrate dieta...

    2024

  7. [15]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “Nerf: Representing scenes as neural radiance fields for view synthesis,” Communications of the ACM , vol. 65, no. 1, pp. 99–106, 2021

    2021

  8. [16]

    3d gaussian splatting for real-time radiance field rendering

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering.” ACM Trans. Graph. , vol. 42, no. 4, pp. 139–1, 2023

    2023

  9. [17]

    3d reconstruction and volume estimation of food using stereo vision techniques,

    F. Konstantakopoulos, E. I. Georga, and D. I. Fotiadis, “3d reconstruction and volume estimation of food using stereo vision techniques,” in 2021 IEEE 21st International Conference on Bioinformatics and Bioengineer- ing (BIBE). IEEE, 2021, pp. 1–4

    2021

  10. [18]

    3dgsr: Implicit surface reconstruction with 3d gaussian splatting,

    X. Lyu, Y .-T. Sun, Y .-H. Huang, X. Wu, Z. Yang, Y . Chen, J. Pang, and X. Qi, “3dgsr: Implicit surface reconstruction with 3d gaussian splatting,” ACM Transactions on Graphics (TOG) , vol. 43, no. 6, pp. 1–12, 2024

    2024

  11. [19]

    Pre-nerf 360: Enriching unbounded appearances for neural radiance fields,

    A. AlMughrabi, U. Haroon, R. Marques, and P. Radeva, “Pre-nerf 360: Enriching unbounded appearances for neural radiance fields,” arXiv preprint arXiv:2303.12234, 2023

    2023 arXiv

  12. [20]

    Structure-from-motion revisited,

    J. L. Sch ¨onberger and J.-M. Frahm, “Structure-from-motion revisited,” in Conference on Computer Vision and Pattern Recognition (CVPR) , 2016

    2016

  13. [21]

    Two- view 3d reconstruction for food volume estimation,

    J. Dehais, M. Anthimopoulos, S. Shevchik, and S. Mougiakakou, “Two- view 3d reconstruction for food volume estimation,” IEEE transactions on multimedia, vol. 19, no. 5, pp. 1090–1099, 2016

    2016

  14. [22]

    Image- based food volume estimation,

    C. Xu, Y . He, N. Khannan, A. Parra, C. Boushey, and E. Delp, “Image- based food volume estimation,” in Proceedings of the 5th international workshop on Multimedia for cooking & eating activities , 2013, pp. 75– 80

    2013

  15. [23]

    Diabetes60- inferring bread units from food images using fully convolutional neural networks,

    P. Ferdinand Christ, S. Schlecht, F. Ettlinger, F. Grun, C. Heinle, S. Tatavatry, S.-A. Ahmadi, K. Diepold, and B. H. Menze, “Diabetes60- inferring bread units from food images using fully convolutional neural networks,” in Proceedings of the IEEE International Conference on C...

    2017

  16. [24]

    Im2calories: towards an automated mobile vision food diary,

    A. Meyers, N. Johnston, V . Rathod, A. Korattikara, A. Gorban, N. Sil- berman, S. Guadarrama, G. Papandreou, J. Huang, and K. P. Murphy, “Im2calories: towards an automated mobile vision food diary,” in Proceedings of the IEEE international conference on computer vision , 2015,...

    2015

  17. [25]

    Nutrition5k: Towards automatic nutritional understanding of generic food,

    Q. Thames, A. Karpur, W. Norris, F. Xia, L. Panait, T. Weyand, and J. Sim, “Nutrition5k: Towards automatic nutritional understanding of generic food,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 8903–8911

    2021

  18. [26]

    Low texture 3d reconstruction based on deep learning hierarchical networks,

    Y . Hou, Z. Yang, Y . Hu, and L. Wang, “Low texture 3d reconstruction based on deep learning hierarchical networks,” in 2024 2nd Interna- tional Conference on Algorithm, Image Processing and Machine Vision (AIPMV). IEEE, 2024, pp. 176–179

    2024

  19. [27]

    Arcore overview,

    Google Developers, “Arcore overview,” https://developers.google.com/ ar/develop/, 2024, accessed: 2024-12-10

    2024

  20. [28]

    Arkit - augmented reality,

    Apple Developer, “Arkit - augmented reality,” https://developer.apple. com/augmented-reality/arkit/, 2024, accessed: 2024-12-10

    2024

  21. [29]

    Instant neural graphics primitives with a multiresolution hash encoding,

    T. M ¨uller, A. Evans, C. Schied, and A. Keller, “Instant neural graphics primitives with a multiresolution hash encoding,” ACM Trans. Graph. , vol. 41, no. 4, pp. 102:1–102:15, Jul. 2022. [Online]. Available: https://doi.org/10.1145/3528223.3530127

    2022

  22. [30]

    Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,

    P. Wang, L. Liu, Y . Liu, C. Theobalt, T. Komura, and W. Wang, “Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,” arXiv preprint arXiv:2106.10689 , 2021

    2021 arXiv

  23. [31]

    Neus2: Fast learning of neural implicit surfaces for multi- view reconstruction,

    Y . Wang, Q. Han, M. Habermann, K. Daniilidis, C. Theobalt, and L. Liu, “Neus2: Fast learning of neural implicit surfaces for multi- view reconstruction,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 3295–3306

    2023

  24. [32]

    Large scale multi-view stereopsis evaluation,

    R. Jensen, A. Dahl, G. V ogiatzis, E. Tola, and H. Aanæs, “Large scale multi-view stereopsis evaluation,” in 2014 IEEE Conference on Computer Vision and Pattern Recognition . IEEE, 2014, pp. 406–413

    2014

  25. [33]

    Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,

    S. Zheng, J. Lu, H. Zhao, X. Zhu, Z. Luo, Y . Wang, Y . Fu, J. Feng, T. Xiang, P. H. Torr et al. , “Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition...

    2021

  26. [34]

    Xmem++: Production-level video segmentation from few annotated frames,

    M. Bekuzarov, A. Bermudez, J.-Y . Lee, and H. Li, “Xmem++: Production-level video segmentation from few annotated frames,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 635–644

    2023

  27. [35]

    Food volume estimation for quantifying dietary intake with a wearable camera,

    A. Gao, F. P.-W. Lo, and B. Lo, “Food volume estimation for quantifying dietary intake with a wearable camera,” in 2018 IEEE 15th International Conference on Wearable and Implantable Body Sensor Networks (BSN) . IEEE, 2018, pp. 110–113

    2018

  28. [36]

    Distinctive image features from scale-invariant keypoints,

    D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” International journal of computer vision , vol. 60, pp. 91–110, 2004. 13

    2004

  29. [37]

    Local invariant feature detectors: a survey,

    T. Tuytelaars, K. Mikolajczyk et al., “Local invariant feature detectors: a survey,” Foundations and trends® in computer graphics and vision , vol. 3, no. 3, pp. 177–280, 2008

    2008

  30. [38]

    Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,

    M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,” Communications of the ACM, vol. 24, no. 6, pp. 381–395, 1981

    1981

  31. [39]

    Openmvs: Open multiple view stereovision,

    D. Cernea, “Openmvs: Open multiple view stereovision,” 2015. [Online]. Available: https://github.com/cdcseacave/openMVS/

    2015

  32. [40]

    Mvs- boost: An efficient point cloud-based 3d reconstruction,

    U. Haroon, A. AlMughrabi, R. Marques, and P. Radeva, “Mvs- boost: An efficient point cloud-based 3d reconstruction,” arXiv preprint arXiv:2406.13515, 2024

    2024 arXiv

  33. [41]

    Divergence theorem,

    Various, “Divergence theorem,” Accessed: March 19, 2025, wikipedia article. [Online]. Available: https://en.wikipedia.org/wiki/Divergence theorem

    2025

  34. [42]

    Neus2: Fast learning of neural implicit surfaces for multi- view reconstruction,

    Y . Wang, Q. Han, M. Habermann, K. Daniilidis, C. Theobalt, and L. Liu, “Neus2: Fast learning of neural implicit surfaces for multi- view reconstruction,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , 2023

    2023

  35. [43]

    Human performance modeling and rendering via neural animated mesh,

    F. Zhao, Y . Jiang, K. Yao, J. Zhang, L. Wang, H. Dai, Y . Zhong, Y . Zhang, M. Wu, L. Xu et al. , “Human performance modeling and rendering via neural animated mesh,” ACM Transactions on Graphics (TOG), vol. 41, no. 6, pp. 1–17, 2022

    2022

  36. [44]

    Parametric correspondence and chamfer matching: Two new techniques for image matching,

    H. G. Barrow, J. M. Tenenbaum, R. C. Bolles, and H. C. Wolf, “Parametric correspondence and chamfer matching: Two new techniques for image matching,” in Proceedings: Image Understanding Workshop . Science Applications, Inc, 1977, pp. 21–27

    1977

  37. [45]

    Displacement (fluid),

    Various, “Displacement (fluid),” Accessed: March 19, 2025, wikipedia article. [Online]. Available: https://en.wikipedia.org/wiki/Displacement (fluid)

    2025

Pith tools

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