Pith. sign in

REVIEW 3 major objections 6 minor 36 references

Lightweight Multiplane Images Network for Real-Time Stereoscopic Conversion from Planar Video

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read A 26M-parameter multiplane-image network converts planar video to stereo at 2K resolution in 18 ms per frame, over 40 times faster than a state-of-the-art temporal-MPI method while matching its perceptual quality.

desk verdict Solid lightweight MPI stereo conversion; the 40x speedup is plausible, but the low-res rendering assumption needs a GT-based check before the real-time claim fully lands. read the letter →

arxiv 2412.03102 v1 pith:VM2KEMBB submitted 2024-12-04 cs.CV

classification cs.CV
keywords planar-to-stereoconversionmultiplaneimagesreal-time3Dvideolightweightnetworkdepth-awarefeatures2Kviewsynthesis
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 tries to establish that a small multiplane-image network with 26 million parameters can convert ordinary planar video into stereoscopic video at 2K resolution in real time, without needing an explicit depth map at inference. It argues that a lightweight depth-semantic branch can learn depth-aware features under supervision from a heavy pretrained monocular depth estimator during training, and that the expensive MPI rendering can be computed at low resolution and then upsampled. If correct, this would make MPI-based 3D conversion practical for glasses-free displays and VR, where current high-quality methods take hundreds of milliseconds per frame. The reported result is 18 ms per 2K frame versus 721 ms for the temporal MPI baseline, with nearly equal perceptual scores.

What carries the argument

The central object is the multiplane image (MPI): a scene represented as N fronto-parallel RGBA planes at fixed depths, rendered into a novel view by horizontal pixel warping followed by back-to-front alpha compositing (the 'over' operation). The paper's mechanism is to generate these planes with two encoders: a detail branch that encodes the full-resolution texture once per frame, and a lightweight depth-semantic branch with a large receptive field that produces depth-aware features; mask heads slice those features into per-plane masks, and a small decoder produces the color-plus-density planes. A training-only auxiliary branch, consisting of two depth heads and a pretrained monocular depth reference, forces the depth-semantic features to align with real depth, and the low-resolution rendering shortcut in Eq. (9) replaces full-resolution plane generation with bilinear upsampling.

What would settle it

Test the method on a suite of 2K videos with large foreground objects and textured backgrounds, such as a person walking in front of foliage; if the low-resolution render produces visibly blurred or jagged edges in the disoccluded regions, or if its MOS drops noticeably below the full-resolution variant, the load-bearing assumption in Eq. (9) is falsified.

Watch

Extended reading notes

Core claim

The central claim is that a 26M-parameter MPI network with no explicit depth input at inference can match the perceptual quality of a 37M-parameter temporal MPI model plus a 123M-parameter depth estimator while running over 40 times faster (18 ms versus 721 ms per 2K frame). The architecture couples a detail branch that extracts features from the input frame with a depth-semantic branch that encodes coarse depth cues; during training a coarse-to-fine auxiliary branch with two depth heads is supervised by a large pretrained monocular depth network, but that branch is discarded at inference. Rendering is accelerated by computing the 16 MPI planes at reduced resolution and bilinearly upsampling the blend weights and colors before compositing, on the claim that the artifacts of MPI upsampling are not prominent in small, smooth occlusion regions. The paper reports LPIPS 0.036 versus 0.034 for TMPI and comparable blind-quality and MOS scores on its 2K test set.

Load-bearing premise

The real-time and quality claims rest on the assumption that computing the MPI planes at low resolution and bilinearly upsampling them does not visibly harm the rendered stereo view, which the paper justifies by saying that occluded regions are small and smooth and supports with ten Youku2K clips.

Editorial extensions

If this is right

  • Real-time 2K stereoscopic conversion becomes feasible on a single modern GPU, since the reported 18 ms per frame is well under the typical interactive budget for glasses-free 3D displays.
  • MPI methods no longer need an external monocular depth estimator at inference, removing a large parameter dependency and its associated latency.
  • The training-only depth auxiliary branch means the model can be supervised with any pretrained depth estimator without adding any inference cost.
  • The reduction from hundreds of millions of parameters to 26M makes deployment on embedded or mobile devices more plausible.

Reading between the lines

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

  • An unstated extension is that the low-resolution rendering shortcut should transfer to 4K and 8K input with similar speed scaling, since the MPI computation itself is resolution-independent; the 'small and smooth occlusion' assumption is where that transfer would likely strain.
  • The same depth-semantic-branch-plus-training-auxiliary recipe could be applied to free-viewpoint video, not just fixed-baseline stereo, by rendering from a different target viewpoint per frame.
  • Because the depth reference network is needed only during training, the final model could be trained with synthetic depth from graphics engines, decoupling its quality from the reference network's failure modes without affecting inference cost.
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

3 major / 6 minor

Summary. The paper proposes LMPIN, a lightweight MPI-based network for planar-to-stereo video conversion. The network contains a detail branch and a lightweight depth-semantic branch; a training-only auxiliary branch supervises depth perception through a pretrained monocular depth model, allowing inference without explicit depth input. The main acceleration comes from computing the MPI at low resolution and upsampling it with bilinear interpolation while blending with the original high-resolution image (Eq. 9). The authors evaluate on a 3D movie test set with ground truth and on ten 2K YouTube-style videos with no-reference metrics. They report LPIPS 0.036, PSNR 33.037, and SSIM 0.913 on the 3D movie set (TMPI: 0.034, 33.630, 0.924), and 18 ms per frame at 2K versus 721 ms for TMPI, claiming over 40x acceleration with similar subjective quality.

Significance. If the results hold, the paper offers a practical real-time 2D-to-3D conversion solution at 2K resolution, a meaningful step for glasses-free 3D displays and VR. The architecture is compact (26M parameters) and effectively combines a lightweight depth-semantic branch with a training-only heavy supervision branch. The experiments include several baselines, ablations of the network components, and an ablation of the low-resolution rendering. However, the central real-time/quality claim rests on an empirical assumption in Eq. (9) that is only validated with no-reference metrics on a small set of clips, and the paper omits the downsampling factor and hardware details needed to reproduce the runtime. These issues must be addressed before the claims can be fully accepted.

major comments (3)
  1. [Accelerate Rendering with Low-Resolution MPI, Eq. (9); Table 4] The central real-time speed-up relies on the assumption in Eq. (9) that computing MPI planes at low resolution and upsampling them does not visibly degrade the synthesized view. The only evidence provided is Table 4, which uses no-reference quality metrics (MUSIQ, HIQA, NIQE) on ten Youku2K clips; the 3D movie test set, which has ground truth, is never used to compare the low-resolution rendering against the full-resolution variant. Since the proposed method is already 0.6 dB below TMPI in PSNR on the 3D movie test set (Table 1), the paper should report PSNR/SSIM/LPIPS for the full-resolution variant on that set to show whether Eq. (9) contributes to the gap. This experiment is load-bearing: if Eq. (9) degrades quality on videos with large or textured disocclusions, the real-time claim would require full-resolution MPI at 155 ms per frame (Table 4), eliminating the headline speed advantage.
  2. [Method, Accelerate Rendering with Low-Resolution MPI] The downsampling factor for the low-resolution MPI is never specified. The text states only that the MPI is computed 'in low-resolution space,' but the resolution of the planes (e.g., 480×270 for 2K input) is required to reproduce the 18 ms runtime and to judge the trade-off. Please state the factor and, ideally, ablate over it on a ground-truth benchmark, since this free parameter directly controls the speed/quality balance.
  3. [Experiments, Datasets and Implementation Details] The runtime comparison in Table 2 is reported without specifying the GPU model or measurement protocol. Since real-time inference is a headline contribution, the hardware (e.g., NVIDIA V100 vs RTX 3090) and the measurement method (e.g., average over frames after warm-up) must be stated. Without this, the 'over 40x acceleration' number is not reproducible and the comparison across methods may be affected by different implementation efficiencies.
minor comments (6)
  1. [General / typos] There are several typos, e.g., 'Experimential Results' in the Experiments section and 'Full resotion' in Table 4; these should be corrected.
  2. [Eq. (8)] In the definition of w_n, the product over j>n uses α_n_s instead of α_j_s; the subscript should be j to match the index of the product.
  3. [Table 4] The 'Bilinear interpolation' row is not clearly defined; specify whether it upscales the final rendered image or the MPI planes, and describe how it differs from the proposed low-resolution rendering.
  4. [Table 2 / MOS] The MOS results are based on 15 observers with no reported variance or significance test; please state the number of evaluated clips, the display device, and provide confidence intervals, since the 0.05 difference between the proposed method and TMPI is within likely inter-observer noise.
  5. [Method, Detail Branch] The detail branch is described only as 'a commonly used autoencoder (Zhang et al. 2023)'; the exact architecture or a precise reference should be given for reproducibility.
  6. [Experiments, Test set] The resolution of the 3D movie test set is not stated in Table 1; if it is different from the 2K Youku test set, this should be explicitly mentioned so the reader can interpret the two evaluations separately.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the method is trained with external supervision, evaluated on external benchmarks, and its low-resolution rendering assumption is an independently testable approximation.

full rationale

The paper's central claims are an architecture design and a low-resolution MPI rendering approximation. The network is trained with external supervision (DPT reference depth maps and ground-truth target views) and evaluated on an external 3D movie test set used by prior work, plus the Youku2K set for high-resolution testing. The authors' prior TMPI work appears only as a comparison baseline, not as a premise in the derivation: no equation, loss term, or architectural choice is forced by TMPI, and the proposed method does not inherit TMPI's temporal mechanism or its outputs. The low-resolution rendering assumption in Eq. 9 is an independent, falsifiable approximation with its own ablation study in Table 4; any weakness in its empirical validation is a correctness or robustness concern, not circularity. The paper does not fit a parameter to a target result and then rename it a prediction, nor does it invoke a uniqueness theorem, and it does not smuggle an ansatz through a self-citation. The use of DPT depth only during training is explicit, and inference neither requires nor claims to reproduce DPT. Overall, the derivation is self-contained and its claims are checked against external data, so no circular step is identified.

Assumptions & free parameters 3 free parameters · 4 assumptions · 2 invented entities

The central claims rest on a few tuned design choices (N=16, lambda=10, low-resolution scale) and on four domain assumptions: MPI sufficiency, DPT reliability as a teacher, synthetic data validity, and the low-res rendering assumption. The paper introduces two new modules (DEFF and the auxiliary branch) whose evidence is internal ablation only. No code, checkpoints, or datasets are released to independently verify these choices.

free parameters (3)
  • number of MPI planes N = 16
    Chosen because disparity between left and right views is small in stereo video; no systematic sweep is reported, and an ablation with 64 planes degrades scores.
  • mask loss weight lambda = 10
    Set experimentally so depth losses and mask loss are on the same order of magnitude; affects training of the depth-semantic branch.
  • MPI low-resolution factor = not specified
    The paper states MPI is computed in low-resolution space but does not state the downsampling factor, which directly controls the speed-up and quality trade-off.
assumptions (4)
  • domain assumption Fronto-parallel MPI with uniformly spaced disparity planes is a sufficient scene representation for stereo conversion from frontal views.
    Stated in the Related Work and Method sections; if scenes have large non-frontal geometry, MPI warping would produce artifacts.
  • domain assumption DPT depth estimates are accurate enough to serve as pseudo ground truth for training the depth-semantic branch.
    The depth losses and mask loss in Eq. 10-12 supervise against DPT's predicted depth; any systematic DPT bias becomes a training bias.
  • domain assumption Synthetic stereo pairs generated from COCO via estimated disparity (Watson et al. method) approximate real stereo content well enough to train a generalizable converter.
    The training set is synthetic; the paper does not validate the quality of these pairs or show they match real stereo distributions.
  • domain assumption The simplified low-resolution MPI rendering (Eq. 9) preserves visual quality because occlusion regions are small and smooth.
    This assumption is load-bearing for the real-time claim and directly affects output quality; it is recommended as the weakest assumption.
invented entities (2)
  • Depth-guided enhanced feature fusion (DEFF) block
    purpose: Uses upsampled coarse depth features as attention to fuse detail features from the detail branch, producing fine depth-semantic features.
    The module is new to this paper; its only support is the internal ablation in Table 3, which shows a small drop when replaced by concatenation.
  • Depth-aware auxiliary branch (training-only)
    purpose: Provides coarse-to-fine depth supervision from DPT during training and is removed at inference.
    It is a training mechanism introduced here; support is the ablation and the qualitative depth maps in Fig. 5, without external replication.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lightweight Multiplane Images Network for Real-Time Stereoscopic Conversion from Planar Video." pith.science (2026). https://pith.science/paper/VM2KEMBB

@misc{pith2026241203102,
  author       = {Pith},
  title        = {Pith review of: Lightweight Multiplane Images Network for Real-Time Stereoscopic Conversion from Planar Video},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VM2KEMBB}},
  note         = {Machine review of arXiv:2412.03102}
}
abstract

With the rapid development of stereoscopic display technologies, especially glasses-free 3D screens, and virtual reality devices, stereoscopic conversion has become an important task to address the lack of high-quality stereoscopic image and video resources. Current stereoscopic conversion algorithms typically struggle to balance reconstruction performance and inference efficiency. This paper proposes a planar video real-time stereoscopic conversion network based on multi-plane images (MPI), which consists of a detail branch for generating MPI and a depth-semantic branch for perceiving depth information. Unlike models that depend on explicit depth map inputs, the proposed method employs a lightweight depth-semantic branch to extract depth-aware features implicitly. To optimize the lightweight branch, a heavy training but light inference strategy is adopted, which involves designing a coarse-to-fine auxiliary branch that is only used during the training stage. In addition, the proposed method simplifies the MPI rendering process for stereoscopic conversion scenarios to further accelerate the inference. Experimental results demonstrate that the proposed method can achieve comparable performance to some state-of-the-art (SOTA) models and support real-time inference at 2K resolution. Compared to the SOTA TMPI algorithm, the proposed method obtains similar subjective quality while achieving over $40\times$ inference acceleration.

Figures

Figures reproduced from arXiv: 2412.03102 by the authors.

Figure 1
Figure 1. The proposed lightweight multiplane images stereoscopic conversion network, (a) overall framework, (b) input image, [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Architecture of the proposed lightweight multiplane images stereoscopic conversion network. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The structure of different blocks. (a) depth head, [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The planar-to-stereo conversion results of different methods on 3D movie test set. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 6
Figure 6. Figure 6: LPIPS perception quality and runtime planes of [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: The planar-to-stereo conversion results of different methods for 2K planar videos. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 21 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Caesar, H.; Uijlings, J.; and Ferrari, V. 2018. Coco-stuff: Thing and stuff classes in context. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1209--1218

  4. [4]

    Cheng, H.; Zhang, M.; and Shi, J. Q. 2023. A survey on deep neural network pruning-taxonomy, comparison, analysis, and recommendations. arXiv preprint arXiv:2308.06767

  5. [5]

    Diao, S.; Chen, Y.; Zhao, Y.; Jia, W.; Zhang, Z.; and Wang, R. 2024. Stereo Vision Conversion from Planar Videos Based on Temporal Multiplane Images. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 1519--1527

  6. [6]

    Fehn, C. 2004. Depth-image-based rendering (DIBR), compression, and transmission for a new approach on 3D-TV. In Stereoscopic displays and virtual reality systems XI, volume 5291, 93--104. SPIE

  7. [7]

    Han, Y.; Wang, R.; and Yang, J. 2022. Single-view view synthesis in the wild with learned adaptive multiplane images. In ACM SIGGRAPH 2022 Conference Proceedings, 1--8

  8. [8]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

Show all 36 references
  1. [9]

    Hong, Y.; Zhang, K.; Gu, J.; Bi, S.; Zhou, Y.; Liu, D.; Liu, F.; Sunkavalli, K.; Bui, T.; and Tan, H. 2023. Lrm: Large reconstruction model for single image to 3d. arXiv preprint arXiv:2311.04400

  2. [10]

    Hua, J.; Qiao, W.; and Chen, L. 2022. Recent advances in planar optics-based glasses-free 3D displays. Frontiers in Nanotechnology, 4: 829011

  3. [11]

    Ke, J.; Wang, Q.; Wang, Y.; Milanfar, P.; and Yang, F. 2021. Musiq: Multi-scale image quality transformer. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 5148--5157

  4. [12]

    Kerbl, B.; Kopanas, G.; Leimk \"u hler, T.; and Drettakis, G. 2023. 3D Gaussian Splatting for Real-Time Radiance Field Rendering. ACM Trans. Graph., 42(4): 139--1

  5. [13]

    Lee, J.; Jung, H.; Kim, Y.; and Sohn, K. 2017. Automatic 2d-to-3d conversion using multi-scale deep neural network. In 2017 IEEE International Conference on Image Processing (ICIP), 730--734. IEEE

  6. [14]

    Li, J.; Feng, Z.; She, Q.; Ding, H.; Wang, C.; and Lee, G. H. 2021. Mine: Towards continuous depth mpi with nerf for novel view synthesis. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 12578--12588

  7. [15]

    Liu, G.; Si, J.; Hu, Y.; and Li, S. 2018. Photographic image synthesis with improved U-net. In 2018 Tenth International Conference on Advanced Computational Intelligence, 402--407. IEEE

  8. [16]

    P.; Tancik, M.; Barron, J

    Mildenhall, B.; Srinivasan, P. P.; Tancik, M.; Barron, J. T.; Ramamoorthi, R.; and Ng, R. 2021. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM, 65(1): 99--106

  9. [17]

    completely blind

    Mittal, A.; Soundararajan, R.; and Bovik, A. C. 2012. Making a “completely blind” image quality analyzer. IEEE Signal processing letters, 20(3): 209--212

  10. [18]

    Porter, T.; and Duff, T. 1984. Compositing digital images. ACM SIGGRAPH Computer Graphics, 253–259

  11. [19]

    Ranftl, R.; Bochkovskiy, A.; and Koltun, V. 2021. Vision transformers for dense prediction. In Proceedings of the IEEE/CVF international conference on computer vision, 12179--12188

  12. [20]

    Read, J. 2022. Stereo Vision, Models of. In Encyclopedia of Computational Neuroscience, 3311--3319. Springer

  13. [21]

    Shih, M.-L.; Su, S.-Y.; Kopf, J.; and Huang, J.-B. 2020. 3d photography using context-aware layered depth inpainting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 8028--8038

  14. [22]

    Song, J.; and Lee, S. J. 2023. Knowledge distillation of multi-scale dense prediction transformer for self-supervised depth estimation. Scientific Reports, 13(1): 18939

  15. [23]

    Steffen, L.; Reichard, D.; Weinland, J.; Kaiser, J.; Roennau, A.; and Dillmann, R. 2019. Neuromorphic stereo vision: A survey of bio-inspired sensors and algorithms. Frontiers in neurorobotics, 13: 28

  16. [24]

    Su, S.; Yan, Q.; Zhu, Y.; Zhang, C.; Ge, X.; Sun, J.; and Zhang, Y. 2020. Blindly assess image quality in the wild guided by a self-adaptive hyper network. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 3667--3676

  17. [25]

    Tucker, R.; and Snavely, N. 2020. Single-view view synthesis with multiplane images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 551--560

  18. [26]

    Tulsiani, S.; Tucker, R.; and Snavely, N. 2018. Layer-structured 3d scene inference via view synthesis. In Proceedings of the European Conference on Computer Vision, 302--317

  19. [27]

    M.; Turmukhambetov, D.; Brostow, G

    Watson, J.; Aodha, O. M.; Turmukhambetov, D.; Brostow, G. J.; and Firman, M. 2020. Learning stereo from single images. In Computer Vision--ECCV 2020: 16th European Conference, 722--740. Springer

  20. [28]

    Wofk, D.; Ma, F.; Yang, T.-J.; Karaman, S.; and Sze, V. 2019. Fastdepth: Fast monocular depth estimation on embedded systems. In 2019 International Conference on Robotics and Automation (ICRA), 6101--6108. IEEE

  21. [29]

    Xie, J.; Girshick, R.; and Farhadi, A. 2016. Deep3d: Fully automatic 2d-to-3d video conversion with deep convolutional neural networks. In Computer Vision--ECCV 2016: 14th European Conference, 842--857. Springer

  22. [30]

    Youku . 2019. Youku Video Super-Resolution and Enhancement Challenge dataset (Youku-VSRE2019). https://tianchi.aliyun.com/dataset/39568. Accessed: 2019-09-24

  23. [31]

    Yu, C.; Gao, C.; Wang, J.; Yu, G.; Shen, C.; and Sang, N. 2021. Bisenet v2: Bilateral network with guided aggregation for real-time semantic segmentation. International Journal of Computer Vision, 129: 3051--3068

  24. [32]

    Yu, C.; Wang, J.; Peng, C.; Gao, C.; Yu, G.; and Sang, N. 2018. Bisenet: Bilateral segmentation network for real-time semantic segmentation. In Proceedings of the European conference on computer vision (ECCV), 325--341

  25. [33]

    Zhang, N.; Nex, F.; Vosselman, G.; and Kerle, N. 2023. Lite-mono: A lightweight cnn and transformer architecture for self-supervised monocular depth estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 18537--18546

  26. [34]

    A.; Shechtman, E.; and Wang, O

    Zhang, R.; Isola, P.; Efros, A. A.; Shechtman, E.; and Wang, O. 2018. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition, 586--595

  27. [35]

    Zhang, Z.; and Wang, R. 2022. Temporal3d: 2d-to-3d video conversion network with multi-frame fusion. In 2022 4th International Conference on Advances in Computer Technology, Information Science and Communications (CTISC), 1--5. IEEE

  28. [36]

    Zhou, T.; Tucker, R.; Flynn, J.; Fyffe, G.; and Snavely, N. 2018. Stereo magnification. ACM Transactions on Graphics, 37(4): 1--12

Pith tools

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