Pith. sign in

REVIEW 4 major objections 5 minor 72 references

Error-bounded Point Cloud Compression Using Truncated Octahedron Quantization

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read This paper claims that replacing cubic quantization with truncated-octahedron cells—the most MSE-efficient space-filling polyhedron—lets error-bounded point cloud compressors hit the same distortion with 8.4% fewer voxels and up to 3x highe

desk verdict Solid systems paper whose main theoretical claim is overstated; the geometric optimality of TO is real, but the bitrate bridge is asserted, not proven—still worth reviewing. read the letter →

arxiv 2608.00495 v1 pith:KUK65XKE submitted 2026-08-01 cs.PF cs.DBcs.DC

Youyuan Liu , Longtao Zhang , Ruoyu Li , Bo Jiang , Taolue Yang , Kai Zhao , Sheng Di , Eduard Dragut
show 1 more author
Sian Jin
This is my paper · ORCID
classification cs.PFcs.DBcs.DC MSC 68P3094A2994A34
keywords pointcloudcompressionerror-boundedlossylatticequantizationtruncatedoctahedronBCCspace-fillingcurvesrate-distortionscientificdata
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 argues that nearly all error-bounded point cloud compressors use cube-shaped quantization cells, and that this is geometrically the wrong choice when fidelity is measured by mean squared error. It proves that the truncated octahedron, the Voronoi cell of the body-centered cubic lattice, has the lowest dimensionless second moment among convex space-filling polyhedra, so it achieves the same distortion with 8.4% fewer voxels. The paper then builds XnYZip, a compressor that instantiates this geometry with constant-time BCC quantization, Hilbert space-filling curve serialization, and run-length encoding, and shows on eight datasets that it reaches up to 3x higher compression ratios, 2.2x faster compression, and 1.2x faster decompression than existing compressors at the same distortion. The work matters because dense particle data from scientific simulations is outgrowing storage capacity, and it offers a principled, geometry-driven reason to change the quantization layer rather than merely tune parameters.

What carries the argument

The dimensionless second moment G(P) of a polyhedral voxel—the integrated squared distance to the centroid scaled by volume—carries the comparison of quantization geometries, since lower G(P) at equal volume means lower MSE. The truncated octahedron is the Voronoi cell of the BCC lattice, whose integer coordinates satisfy a simple all-even/all-odd parity rule that makes quantization a constant-time two-candidate nearest-point check. The Hilbert space-filling curve then serializes occupied voxels so that nearby cells stay adjacent in one dimension, converting spatial redundancy into low-entropy first-order differences that entropy coding can exploit.

What would settle it

Measure the post-delta residual entropy H(R) of truncated-octahedron quantization against cube quantization at matched MSE on a dense dataset such as USGS or HACC; the predicted saving is roughly log2(1/0.916) ≈ 0.12 bits per point. Finding H_TO ≥ H_cube would falsify the claim that the geometric voxel-count advantage becomes a rate advantage.

Watch

Extended reading notes

Core claim

Under the high-resolution quantization assumption, the paper establishes a chain of results: an MSE-optimal quantizer's cells must be convex polyhedra; among such cells, efficiency is measured by the dimensionless second moment G(P); and a uniform tiling of a single optimal shape beats any composite tiling. Consulting classical lattice values, the truncated octahedron (Voronoi cell of the BCC lattice) attains the lowest G(P) among space-filling polyhedra, so replacing cubic quantization with truncated-octahedron quantization yields the same MSE with 8.4% fewer occupied voxels. XnYZip realizes this advantage through a parity-based constant-time BCC quantizer, a Hilbert space-filling curve for

Load-bearing premise

The unproved bridge is that minimizing the number of occupied voxels at a fixed error bound always lowers the achievable bitrate; if a coarser truncated-octahedron lattice raises the entropy of the encoded symbol stream enough, the 8.4% geometric voxel saving would not translate into rate savings.

Editorial extensions

If this is right

  • In dense regimes (occupancy rho much greater than 1), the truncated-octahedron lattice delivers a stable voxel-count reduction of about 8.6% compared with cubes at matched error bounds, and the advantage shrinks smoothly as occupancy approaches 1.
  • XnYZip sits at or above the Pareto frontier on all eight evaluated datasets and dominates existing point cloud compressors in the dense, low-PSNR regime predicted by the theory.
  • The geometric gain costs little wall-clock time: switching the quantization lattice from cube to truncated octahedron adds only about 3–16% end-to-end overhead while materially improving the compression ratio.
  • The chunked compressed representation doubles as a spatial histogram and region-aggregation index, speeding radius-count queries by 7.5–34.6x compared with decode-then-filter while remaining bit-exact.
  • The central geometric advantage is density-dependent, so the same compressor has limited benefit on sparse point clouds where occupancy approaches one point per voxel.

Reading between the lines

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

  • The same dimensionless-second-moment argument would extend to higher-dimensional lattice quantizers—such as the E8 lattice in eight dimensions—for rate-distortion in vector-embedding compression, though the paper only treats 3D point clouds.
  • A direct measurement of post-delta residual entropy for TO versus cube quantization at matched MSE would isolate the geometric gain from the encoder's contribution; the paper reports entropy for different space-filling curves but not this specific lattice comparison.
  • The paper's theory is MSE-specific, but its own Table 3 suggests the truncated octahedron also has better Hausdorff covering efficiency, hinting that a similar geometry-first argument could be made for maximum-error-bounded compression, a direction the paper does not develop.
  • An adaptive hybrid that uses cubes in sparse regions and truncated octahedra in dense regions could realize the geometric gain on mixed-distribution datasets; the paper lists adaptive hybrid quantization as future work rather than evaluating it.
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 argues that existing error-bounded point-cloud compressors are suboptimal because they quantize with a cubic lattice, whereas the truncated octahedron (TO), the Voronoi cell of the BCC lattice, has the lowest dimensionless second moment among space-filling polyhedra and therefore minimizes MSE at fixed quantizer resolution. The authors develop XnYZip, a compressor that replaces cube quantization with TO quantization, serializes the resulting voxel indices with a space-filling curve, and applies delta/RLE/Huffman/Zstd coding. They report up to 3× higher compression ratios, 2.2× faster compression, and 1.2× faster decompression than state-of-the-art baselines on dense scientific point clouds, together with a distributed MPI implementation and a sidecar-based query path. The central theoretical bridge is the claim in Section 3 that minimizing the number of occupied voxels M implies minimizing the achievable rate R under the proposed locality-preserving encoder.

Significance. If the central claim holds, the paper would make a useful contribution: it connects classical lattice quantization theory (Conway–Sloane, Gersho) to practical point-cloud compression, gives a concrete geometric design principle, and provides an extensive evaluation with matched-error comparisons, statistical significance tests, KL-divergence checks of the uniformity assumption, and a reproducible artifact. The 8.4%/8.6% voxel-count advantage is derived from published G(P) values rather than fitted, and the empirical confirmation of the density-dependent behavior is a strength. However, the manuscript's main theoretical assertion—that the geometric voxel reduction translates into bitrate savings—is not proven, and the evaluation does not isolate the TO contribution from the rest of the pipeline. These gaps are load-bearing for the paper's core claim and require a major revision.

major comments (4)
  1. [Section 3, 'min( M ) ⇒ min( R )'] The bridge from minimizing occupied voxel count M to minimizing achievable rate R is asserted without derivation. The actual rate of XnYZip is determined by the entropy of the SFC/delta/RLE/Huffman/Zstd stream, which depends on the lattice alphabet size, the occupancy pattern, the count distribution, and the conditional entropy of the residual stream, not simply on M. High-rate quantization theory gives a rate advantage of only (3/2) log2(G_cube/G_TO) ≈ 0.128 bits per point from the second-moment gain, which is much smaller than the 8.4% voxel reduction unless total rates are extremely low. The paper does not provide a controlled ablation in which only the quantizer is toggled (cube vs. TO) while keeping all other pipeline stages fixed; Table 9 reports only timing, not compression ratio. Without such an ablation or a rigorous bound connecting M to the coded length of the specific entropy
  2. [Section 4.1 and Table 7 (theory vs. evaluation criterion)] The theoretical optimality is developed for MSE: the dimensionless second moment G(P) is the appropriate figure of merit when distortion is mean squared error. However, the headline compression-ratio comparisons in Table 7 are reported at matched relative L2 max-error targets (|p−p'|2/diag ≤ target), i.e., a covering-radius / Hausdorff-style criterion. For a fixed maximum-error constraint, the optimal lattice is not necessarily the one minimizing G(P); one should minimize volume for a given covering radius (or, equivalently, maximize covering density). The 8.4% voxel advantage in Table 3 is computed for equal MSE, not equal max-error. The paper does not prove that the BCC lattice is optimal—or even that it retains the same advantage—under the max-error metric used in the main evaluation. Please either derive the corresponding covering-radius theory or report the controlled comparison at
  3. [Table 9 and §5.3] The end-to-end gains reported in Figure 8 and Table 7 could plausibly be driven by the RLE stage, the SFC ordering (Z-order vs. Hilbert), the Huffman dictionary, or the Zstd backend, rather than by the TO geometry itself. The only quantizer-to-quantizer comparison inside the same pipeline is the timing study in Table 9, which does not report compression ratios. A proper ablation should run XnYZip with cube quantization and TO quantization under identical conditions on the same datasets and error targets, showing the contribution of the geometry alone. This ablation is necessary to support the paper's title claim and the 'provably optimal' language. Without it, the empirical results leave the central causal link unverified.
  4. [Lemma 4.3] The proof of Lemma 4.3 is mostly sound, but the statement 'this inequality is strict in any practical composite tiling, as it will either contain suboptimal shapes (including rotated versions of an optimal shape)' is imprecise: rotating a polyhedron does not change its G value, so rotated versions of the optimal shape are not suboptimal. Strictness follows from non-uniform volumes (Jensen's inequality) or from genuinely suboptimal shapes, but not from rotation alone. This should be corrected for rigor.
minor comments (5)
  1. [Abstract and §1] 'up to 3× higher compression ratios' should specify the reference baseline and dataset/error-bound configuration; the abstract currently leaves this ambiguous.
  2. [§5.1, Table 4] The HACC row states 'Tested Size' as 12.9 GB, but the text says a smaller subset (~3.13 GB) is used for compression ratio and case study evaluation. Please clarify which size corresponds to which experiment.
  3. [§4.2] The BCC quantization description says the scale is '2ε/√5', but the derivation of this scale (relationship between the user error bound ε and the BCC lattice spacing) is not given. A short derivation or reference would help the reader verify the error-bound guarantee.
  4. [§5.2, Figure 7] The empirical density advantage is reported as '≈8.6%' but the caption and text do not state the standard deviation or the number of datasets/error points used to compute this average. A confidence interval would strengthen the claim.
  5. [§5.3, Figure 8] The phrase 'distortion rate per–dataset' is a typo; it should read 'rate–distortion per dataset'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the TO optimality claim rests on external lattice-theory constants, the 8.4% voxel reduction is derived algebraically, and the empirical results are measured, not fitted.

full rationale

The paper's central derivation chain is: formulate compression as quantization, prove that MSE-optimal voxels are convex polyhedra, use the dimensionless second moment G(P) as the comparison criterion, then select the space-filling polyhedron with minimal G(P). The G(P) values in Table 2 are taken from the external classical result of Conway and Sloane [10], not from the authors' own prior work, and the 8.4% voxel-count advantage in Table 3 follows algebraically from the ratio of G values at equal MSE. The empirical 8.6% density reduction in Figure 7 is a measured confirmation, not a fitted parameter used to produce the compression-ratio claims. The compression-ratio and throughput improvements are obtained by comparing against independent baselines (SZ3, ZFP, Draco, TMC13, ALP, BUFF, LCP) under matched error bounds; no parameter is fitted to the reported ratios. The paper does contain self-citations to LCP and other same-group systems, but these are used as baselines/background and are not load-bearing for the theoretical optimality argument. The main weakness—the asserted bridge min(M) => min(R) in Section 3—is an unproven monotonicity assumption about rate versus voxel count, and the paper does not report a controlled cube-vs-TO bitrate ablation; however, this is a correctness/evidence concern, not circularity, because the assertion is not defined in terms of the conclusion and the subsequent experimental claims are not derived from it by construction. Hence no circular step can be exhibited, and the appropriate score is 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

No free parameters are fitted to produce the central claim; the BCC lattice scale 2epsilon/sqrt(5) is set by the user error bound. The paper relies on classical lattice-quantization results plus two domain assumptions (spatial uniformity, high-resolution regime) and one asserted modeling bridge (min(M) => min(R)). No invented entities.

assumptions (4)
  • domain assumption Spatial Uniformity Assumption: points are uniformly distributed within each quantization voxel.
    Stated in Section 4.1.1, validated in Figures 4 and 6. Needed for the dimensionless-second-moment G(P) analysis to predict actual MSE behavior.
  • ad hoc to paper min(M) => min(R): the achievable rate is monotonically non-increasing in the number of occupied voxels under a locality-preserving entropy-coded encoder.
    Section 3, 'The Optimal Quantization Problem'. Asserted without proof. This is the bridge from geometric voxel-count savings to bitrate savings, and is the weakest load-bearing premise.
  • domain assumption High-resolution regime: quantization error is much smaller than intrinsic spatial variation of the data.
    Section 4.1.1: needed for the local uniform-distribution approximation; claimed to align with high-accuracy scientific requirements.
  • standard math External theorem: the BCC lattice (TO Voronoi cell) has the lowest dimensionless second moment among 3D lattice quantizers.
    Invoked via Conway-Sloane [10] and Lookabaugh-Gray [48] to conclude TO optimality in Table 2; not proven in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Error-bounded Point Cloud Compression Using Truncated Octahedron Quantization." pith.science (2026). https://pith.science/paper/KUK65XKE

@misc{pith2026260800495,
  author       = {Pith},
  title        = {Pith review of: Error-bounded Point Cloud Compression Using Truncated Octahedron Quantization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KUK65XKE}},
  note         = {Machine review of arXiv:2608.00495}
}
read the original abstract

With the rapid advancement of large-scale scientific simulations, the massive volume of point cloud data generated has increasingly become a critical bottleneck for scientific storage systems and data management pipelines. Existing point cloud compression techniques integrated into scientific storage systems are designed for sparse geometry and rely on quantization schemes whose optimality assumptions do not hold for dense data. When applied at the compression layer to point clouds, this representation mismatch leads to fundamentally sub-optimal rate-distortion trade-offs that cannot be addressed through parameter tuning or framework-level adaptations. This mismatch increases storage overhead and limits efficient movement and downstream analysis of simulation outputs. This issue arises in scientific data management workflows handling large-scale dense particle datasets. State-of-the-art compression methods fail to fully exploit the redundancies inherent in such data. We address this limitation by developing a theory of point cloud compressibility for dense data, characterizing fundamental rate-distortion behavior at the representation layer. Guided by this analysis, we introduce XnYZip, an error-bounded lossy compressor based on provably optimal Truncated Octahedron quantization, combined with a locality-aware encoding pipeline using space-filling curves and run-length encoding. Experiments on large-scale scientific datasets demonstrate consistent storage and throughput improvements, achieving up to 3x higher compression ratios, 2.2x faster compression, and 1.2x faster decompression compared to state-of-the-art point cloud compressors under same distortion.

Figures

Figures reproduced from arXiv: 2608.00495 by the authors.

Figure 1
Figure 1. A visual comparison of quantization schemes using the [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Visual comparison of three traversal paths on a [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The overview of XnYZip. (a) Cube quant, USGS dataset. (b) TO quant, USGS dataset. (c) Cube quant, Stanford Bunny. (d) TO quant, Stanford Bunny [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Comparison of the actual quantization error distributions [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: The centers of truncated octahedra form a BCC layout. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Kullback-Leibler divergence measuring deviation between [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Geometric efficiency of Truncated Octahedron (TO) quanti [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 9
Figure 9. Figure 9: Critical-difference diagram for the six lossy compressors [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 8
Figure 8. Figure 8: Rate–distortion. XnYZip is at or above the Pareto frontier on every dataset and dominates in the low-PSNR (dense) regime. 5.3 Performance Analysis Next, we evaluate XnYZip’s end-to-end behaviour against the base￾lines on the same eight datasets in different metrics. Ra…
Figure 10
Figure 10. Figure 10: Stanford Dragon reconstructions at a fixed CR of [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 11
Figure 11. Figure 11: Throughput vs. bitrate on USGS and Vesicles. At low bpps [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 12
Figure 12. Figure 12: HACC radial density profile at 𝑒𝑏 = 10−3 (CR ≈ 134×). Left: original particle counts per spherical shell and the reconstructed profile coincide to within line thickness across all 50 shells. Right: per-shell relative error Δ𝑛/𝑛. The mean absolute deviation is 0.41%; t…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

72 extracted references · 58 canonical work pages

  1. [1]

    https://docs.hdfgroup.org/hdf5/develop/_f_i_l_t_e_r.html Online

    2023.HDF5 Filters. https://docs.hdfgroup.org/hdf5/develop/_f_i_l_t_e_r.html Online

    work page 2023

  2. [2]

    Kuffo, and Peter Boncz

    Azim Afroozeh, Leonardo X. Kuffo, and Peter Boncz. 2023. ALP: Adaptive Lossless floating-Point Compression.Proc. ACM Manag. Data1, 4, Article 230 (Dec. 2023), 26 pages. https://doi.org/10.1145/3626717

    work page doi:10.1145/3626717 2023

  3. [3]

    Nitin Agrawal and Ashish Vulimiri. 2017. Low-Latency Analytics on Colos- sal Data Streams with SummaryStore. InProceedings of the 26th Symposium on Operating Systems Principles(Shanghai, China)(SOSP ’17). Association for Computing Machinery, New York, NY, USA, 647–664. https://doi.org/10.1145/ 3132747.3132758

    arXiv 2017

  4. [4]

    Mark Ainsworth, Ozan Tugluk, Ben Whitney, and Scott Klasky. 2018. Multilevel techniques for compression and reduction of scientific data—the univariate case. Computing and Visualization in Science19, 5–6 (2018), 65–76

    work page 2018

  5. [5]

    Chao Cao, Marius Preda, and Titus Zaharia. 2019. 3D Point Cloud Compression: A Survey. InProceedings of the 24th International Conference on 3D Web Technology (LA, CA, USA)(Web3D ’19). Association for Computing Machinery, New York, NY, USA, 1–9. https://doi.org/10.1145/3329714.3338130

    work page arXiv 2019

  6. [6]

    Yufan Chen, Xiangyu Zou, Kaiwen Deng, Hao Hu, Cai Deng, Ke Feng, and Wen Xia. 2025. Apic: A Precomputation-Based Integer Compressor for OLTP Databases. In2025 Data Compression Conference (DCC). 303–312. https://doi. org/10.1109/DCC62719.2025.00038

    work page arXiv 2025

  7. [7]

    Zhiyuan Chen, Johannes Gehrke, and Flip Korn. 2001. Query optimization in compressed database systems. 271–282

    work page 2001

  8. [8]

    Yu Cheng and Florin Rusu. 2014. Parallel in-situ data processing with spec- ulative loading. InProceedings of the 2014 ACM SIGMOD International Con- ference on Management of Data(Snowbird, Utah, USA)(SIGMOD ’14). Asso- ciation for Computing Machinery, New York, NY, USA, 1287–1298. https: //doi.org/10.1145/2588555.2593673

    work page arXiv 2014

Show all 72 references
  1. [9]

    Community Earth System Model (CESM) Atmosphere Model. 2019. http://www. cesm.ucar.edu/models/. Online

    2019

  2. [10]

    Conway and N

    J. Conway and N. Sloane. 1982. Voronoi regions of lattices, second moments of polytopes, and quantization.IEEE Transactions on Information Theory28, 2 (1982), 211–226. https://doi.org/10.1109/TIT.1982.1056483

    1982

  3. [11]

    Cover and Joy A

    Thomas M. Cover and Joy A. Thomas. 2006.Elements of Information Theory (Wiley Series in Telecommunications and Signal Processing). Wiley-Interscience, USA

    2006

  4. [12]

    cuZFP. 2023. https://github.com/LLNL/zfp/tree/develop/src/cuda_zfp. Online

    2023

  5. [13]

    Ismael Daribo, Ryo Furukawa, Ryusuke Sagawa, Hiroshi Kawasaki, Shinsaku Hiura, and Naoki Asada. 2012. Efficient rate-distortion compression of dynamic point cloud for grid-pattern-based 3D scanning systems.3D Research3, 1 (17 Jan 2012), 2. https://doi.org/10.1007/3DRes.01(2012)2

    2012 doi

  6. [14]

    Michael Deering. 1995. Geometry compression. InProceedings of the 22nd Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH ’95). Association for Computing Machinery, New York, NY, USA, 13–20. https: //doi.org/10.1145/218380.218391

    1995

  7. [15]

    2023.H5Z-SZ

    Sheng Di. 2023.H5Z-SZ. https://github.com/disheng222/H5Z-SZ Online

    2023

  8. [16]

    Sheng Di and Franck Cappello. 2016. Fast error-bounded lossy HPC data com- pression with SZ. In2016 IEEE International Parallel and Distributed Processing Symposium. IEEE, IEEE, Chicago, IL, USA, 730–739

    2016

  9. [17]

    EXAALT Project. 2021. EXAALT: Molecular Dynamics at Exascale for Materials Science. https://www.exascaleproject.org/research-project/exaalt/. Online

    2021

  10. [18]

    Haoqiang Fan, Hao Su, and Leonidas Guibas. 2017. A Point Set Generation Network for 3D Object Reconstruction from a Single Image. In2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2463–2471. https://doi.org/10.1109/CVPR.2017.264

    2017 doi

  11. [19]

    Fedeli, A

    L. Fedeli, A. Huebl, F. Boillod-Cerneux, T. Clark, K. Gott, C. Hillairet, S. Jaure, A. Leblanc, R. Lehe, A. Myers, C. Piechurski, M. Sato, N. Zaim, W. Zhang, J. Vay, and H. Vincenti. 2022. Pushing the Frontier in the Design of Laser-Based Electron Accelerators with Groundbreak...

    2022 arXiv

  12. [20]

    Jianyang Gao and Cheng Long. 2024. RaBitQ: Quantizing High-Dimensional Vectors with a Theoretical Error Bound for Approximate Nearest Neighbor Search.Proc. ACM Manag. Data2, 3, Article 167 (May 2024), 27 pages. https: //doi.org/10.1145/3654970

    2024 doi

  13. [21]

    A. Gersho. 1979. Asymptotically optimal block quantization.IEEE Transactions on Information Theory25, 4 (1979), 373–380. https://doi.org/10.1109/TIT.1979. 1056067

    1979 doi

  14. [22]

    Google. 2017. Draco: A library for compressing and decompressing 3D geometric meshes and point clouds. https://github.com/google/draco. Online

    2017

  15. [23]

    Liu, Maria Nieto-Santisteban, Alex Szalay, David J

    Jim Gray, David T. Liu, Maria Nieto-Santisteban, Alex Szalay, David J. DeWitt, and Gerd Heber. 2005. Scientific data management in the coming decade.SIGMOD Rec.34, 4 (Dec. 2005), 34–41. https://doi.org/10.1145/1107499.1107503

    2005

  16. [24]

    Stefan Gumhold, Zachi Kami, Martin Isenburg, and Hans-Peter Seidel. 2005. Pre- dictive point-cloud compression. InACM SIGGRAPH 2005 Sketches(Los Angeles, California)(SIGGRAPH ’05). Association for Computing Machinery, New York, NY, USA, 137–es. https://doi.org/10.1145/1187112.1187277

    2005

  17. [25]

    HACC team (ECP EXASKY). 2019. https://sdrbench.github.io/. Online

    2019

  18. [26]

    David Hilbert. 1891. Ueber die stetige Abbildung einer Line auf ein Flächenstück. Math. Ann.38, 3 (01 Sep 1891), 459–460. https://doi.org/10.1007/BF01199431

    doi

  19. [27]

    Kaisei Hishida, Chunwei Liu, John Paparrizos, and Aaron J. Elmore. 2025. Beyond Compression: A Comprehensive Evaluation of Lossless Floating-Point Compres- sion.Proc. VLDB Endow.18, 11 (July 2025), 4396–4409. https://doi.org/10.14778/ 3749646.3749701

    2025

  20. [29]

    Jay Kuo, and M

    Yan Huang, Jingliang Peng, C.-C. Jay Kuo, and M. Gopi. 2006. Octree-Based Progressive Geometry Coding of Point Clouds. InSymposium on Point-Based Graphics, Mario Botsch, Baoquan Chen, Mark Pauly, and Matthias Zwicker (Eds.). The Eurographics Association. https://doi.org//10.23...

    2006 doi

  21. [30]

    Gowers Richard, Matta Micaela, and Wang Lily

    Alibay Irfan, Beckstein Oliver, Fan Shujie, J. Gowers Richard, Matta Micaela, and Wang Lily. 2018. YiiP equilibrium dataset. https://www.mdanalysis.org/ MDAnalysisData/yiip_equilibrium.html. Online

    2018

  22. [31]

    Iyer and David Wilhite

    Balakrishna R. Iyer and David Wilhite. 1994. Data Compression Support in Databases. InProceedings of the 20th International Conference on Very Large Data Bases (VLDB ’94). Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, 695–704

    1994

  23. [32]

    Søren Kejser Jensen, Torben Bach Pedersen, and Christian Thomsen. 2018. Mod- elarDB: modular model-based time series management with spark and cassandra. Proc. VLDB Endow.11, 11 (July 2018), 1688–1701. https://doi.org/10.14778/ 3236187.3236215

    2018

  24. [33]

    Pu Jiao, Sheng Di, Hanqi Guo, Kai Zhao, Jiannan Tian, Dingwen Tao, Xin Liang, and Franck Cappello. 2022. Toward Quantity-of-Interest Preserving Lossy Com- pression for Scientific Data.Proc. VLDB Endow.16, 4 (Dec. 2022), 697–710. https://doi.org/10.14778/3574245.3574255

    2022

  25. [34]

    Sian Jin, Sheng Di, Jiannan Tian, Suren Byna, Dingwen Tao, and Franck Cappello

  26. [35]

    Joint Photographic Experts Group. 1992. JPEG: Still Image Data Compression Standard (ISO/IEC 10918-1). ITU-T Recommendation T.81 and ISO/IEC 10918-1 Standard. Online; accessed on 2025-10-17

    1992

  27. [36]

    Herve Jégou, Matthijs Douze, and Cordelia Schmid. 2011. Product Quantization for Nearest Neighbor Search.IEEE Transactions on Pattern Analysis and Machine Intelligence33, 1 (2011), 117–128

    2011

  28. [37]

    Julius Kammerl, Nico Blodow, Radu Bogdan Rusu, Suat Gedikli, Michael Beetz, and Eckehard Steinbach. 2012. Real-time compression of point cloud streams. In 2012 IEEE International Conference on Robotics and Automation. 778–785. https: //doi.org/10.1109/ICRA.2012.6224647

    2012

  29. [38]

    Birendra Kathariya, Li Li, Zhu Li, Jose Alvarez, and Jianle Chen. 2018. Scalable Point Cloud Geometry Coding with Binary Tree Embedded Quadtree. In2018 IEEE International Conference on Multimedia and Expo (ICME). 1–6. https: //doi.org/10.1109/ICME.2018.8486481

    2018

  30. [39]

    Kenney and Oliver Beckstein

    Ian M. Kenney and Oliver Beckstein. 2015. Technical Report: SPIDAL Summer REU 2015: Biomolecular benchmark systems. (10 2015). https://doi.org/10.6084/ m9.figshare.1588804.v1

    2015

  31. [40]

    Kersten, Stratos Idreos, Stefan Manegold, and Erietta Liarou

    Martin L. Kersten, Stratos Idreos, Stefan Manegold, and Erietta Liarou. 2011. The researcher’s guide to the data deluge: querying a scientific database in just a few seconds.Proc. VLDB Endow.4, 12 (Aug. 2011), 1474–1477. https: //doi.org/10.14778/3402755.3402799

    2011

  32. [41]

    Masaki Kitago and M. Gopi. 2006. Efficient and Prioritized Point Subsampling for CSRBF Compression. InSymposium on Point-Based Graphics, Mario Botsch, Baoquan Chen, Mark Pauly, and Matthias Zwicker (Eds.). The Eurographics Association. https://doi.org//10.2312/SPBG/SPBG06/121-128

    2006 doi

  33. [42]

    Xin Liang, Sheng Di, Dingwen Tao, Sihuan Li, Shaomeng Li, Hanqi Guo, Zizhong Chen, and Franck Cappello. 2018. Error-controlled lossy compression optimized for high compression ratios of scientific datasets. In2018 IEEE International Conference on Big Data. IEEE, 438–447

    2018

  34. [43]

    Gok, Jiannan Tian, Junjing Deng, Jon C

    Xin Liang, Kai Zhao, Sheng Di, Sihuan Li, Robert Underwood, Ali M. Gok, Jiannan Tian, Junjing Deng, Jon C. Calhoun, Dingwen Tao, Zizhong Chen, and Franck Cappello. 2022. SZ3: A Modular Framework for Composing Prediction-Based Error-Bounded Lossy Compressors.IEEE Transactions o...

    2022

  35. [44]

    Jyh-Ming Lien, Gregorij Kurillo, and Ruzena Bajcsy. 2010. Multi-camera tele- immersion system with real-time model driven data compression.The Visual Computer26, 1 (01 Jan 2010), 3–15. https://doi.org/10.1007/s00371-009-0367-8

    2010 doi

  36. [45]

    Peter Lindstrom. 2014. Fixed-rate compressed floating-point arrays.IEEE Trans- actions on Visualization and Computer Graphics20, 12 (2014), 2674–2683

    2014

  37. [46]

    Chunwei Liu, Hao Jiang, John Paparrizos, and Aaron J. Elmore. 2021. Decomposed bounded floats for fast compression and queries.Proc. VLDB Endow.14, 11 (July 2021), 2586–2598. https://doi.org/10.14778/3476249.3476305

    2021

  38. [47]

    Jinyang Liu, Pu Jiao, Kai Zhao, Xin Liang, Sheng Di, and Franck Cappello. 2025. QPET: A Versatile and Portable Quantity-of-Interest-Preservation Framework for Error-Bounded Lossy Compression.Proc. VLDB Endow.18, 8 (April 2025), 2440–2453. https://doi.org/10.14778/3742728.3742739

    2025

  39. [48]

    Lookabaugh and R.M

    T.D. Lookabaugh and R.M. Gray. 1989. High-resolution quantization theory and the vector quantizer advantage.IEEE Transactions on Information Theory35, 5 (1989), 1020–1033. https://doi.org/10.1109/18.42217

    1989 doi

  40. [49]

    Philipp Merkle, Aljoscha Smolic, Karsten Muller, and Thomas Wiegand. 2007. Multi-View Video Plus Depth Representation and Coding. In2007 IEEE In- ternational Conference on Image Processing, Vol. 1. I – 201–I – 204. https: //doi.org/10.1109/ICIP.2007.4378926

    2007

  41. [50]

    G. M. Morton. 1966.A Computer Oriented Geodetic Data Base; and a New Technique in File Sequencing. Technical Report. IBM Ltd., Ottawa, Canada. Technical Report

    1966

  42. [51]

    MPEG. 2024. MPEG-I Part 9: Geometry-based Point Cloud Compression. https: //www.mpeg.org/standards/MPEG-I/9/. Accessed: 2024-03-22

    2024

  43. [52]

    MPEG Group. 2017. Geometry based point cloud compression (G-PCC) test model. https://github.com/MPEGGroup/mpeg-pcc-tmc13. Online

    2017

  44. [53]

    NYX simulation. 2019. https://amrex-astro.github.io/Nyx/. Online

    2019

  45. [54]

    Tilo Ochotta and Dietmar Saupe. 2004. Compression of point-based 3D models by shape-adaptive wavelet coding of multi-height fields. InProceedings of the First Eurographics Conference on Point-Based Graphics(Switzerland)(SPBG’04). Eurographics Association, Goslar, DEU, 103–112

    2004

  46. [55]

    Meikel Poess and Dmitry Potapov. 2003. Data compression in Oracle. InProceed- ings of the 29th International Conference on Very Large Data Bases - Volume 29 (Berlin, Germany)(VLDB ’03). VLDB Endowment, 937–947

    2003

  47. [57]

    Chou, Robert A

    Sebastian Schwarz, Marius Preda, Vittorio Baroncini, Madhukar Budagavi, Pablo Cesar, Philip A. Chou, Robert A. Cohen, Maja Krivokuća, Sébastien Lasserre, Zhu Li, Joan Llach, Khaled Mammou, Rufael Mekuria, Ohji Nakagami, Ernestasia Siahaan, Ali Tabatabai, Alexis M. Tourapis, an...

    2019

  48. [58]

    Stanford University Computer Graphics Laboratory. 1996. The Stanford 3D Scanning Repository. https://graphics.stanford.edu/data/3Dscanrep/. Online

    1996

  49. [59]

    Dingwen Tao, Sheng Di, Zizhong Chen, and Franck Cappello. 2017. Significantly improving lossy compression for scientific data sets based on multidimensional prediction and error-controlled quantization. In2017 IEEE International Parallel and Distributed Processing Symposium. I...

    2017

  50. [60]

    Gabriel Taubin and Jarek Rossignac. 1998. Geometric compression through topological surgery.ACM Trans. Graph.17, 2 (April 1998), 84–115. https: //doi.org/10.1145/274363.274365

    1998

  51. [61]

    2012.JPEG2000 image compression fundamentals, standards and practice: image compression fundamentals, standards and practice

    David Taubman and Michael Marcellin. 2012.JPEG2000 image compression fundamentals, standards and practice: image compression fundamentals, standards and practice. Vol. 642. Springer Science & Business Media

    2012

  52. [62]

    2023.Hierarchical data format version 5

    The HDF Group. 2023.Hierarchical data format version 5. http://www.hdfgroup. org/HDF5 Online

    2023

  53. [63]

    Hermann Tropf and H. Herzog. 1981. Multimensional Range Search in Dynami- cally Balanced Trees.Angew. Inform.23 (1981), 71–77. https://api.semanticscholar. org/CorpusID:26857103

    1981

  54. [64]

    Geological Survey

    U.S. Geological Survey. 2024. USGS 3DEP LiDAR Point Cloud Dataset. https://www.usgs.gov/news/technical-announcement/usgs-3dep-lidar-point- cloud-now-available-amazon-public-dataset/. Online

    2024

  55. [65]

    Gregory K Wallace. 1992. The JPEG still picture compression standard.IEEE Transactions on Consumer Electronics38, 1 (1992), xviii–xxxiv

    1992

  56. [66]

    Daoce Wang, Jesus Pulido, Pascal Grosset, Sian Jin, Jiannan Tian, James Ahrens, and Dingwen Tao. 2022. TAC: Optimizing Error-Bounded Lossy Compression for Three-Dimensional Adaptive Mesh Refinement Simulations. InProceedings of the 31st International Symposium on High-Performa...

    2022

  57. [67]

    Mingze Xia, Sheng Di, Franck Cappello, Pu Jiao, Kai Zhao, Jinyang Liu, Xuan Wu, Xin Liang, and Hanqi Guo. 2024. Preserving Topological Feature with Sign- of-Determinant Predicates in Lossy Compression: A Case Study of Vector Field Critical Points. In2024 IEEE 40th Internationa...

    2024

  58. [68]

    Mingze Xia, Yuxiao Li, Pu Jiao, Bei Wang, Xin Liang, and Hanqi Guo. 2026. Time- varying Vector Field Compression with Preserved Critical Point Trajectories. arXiv:2510.25143 [cs.DB] https://arxiv.org/abs/2510.25143

    2026

  59. [69]

    Mingze Xia, Bei Wang, Yuxiao Li, Pu Jiao, Xin Liang, and Hanqi Guo. 2025. TspSZ: An Efficient Parallel Error-Bounded Lossy Compressor for Topological Skeleton Preservation . In2025 IEEE 41st International Conference on Data Engineering (ICDE). IEEE Computer Society, Los Alamit...

    2025

  60. [70]

    Horwich, and Paul B

    Zhaohui Xu, Arthur L. Horwich, and Paul B. Sigler. 1997. The crystal structure of the asymmetric GroEL–GroES–(ADP)7 chaperonin complex.Nature388, 6644 (01 Aug 1997), 741–750. https://doi.org/10.1038/41944

    1997 doi

  61. [71]

    Feng Zhang, Zaifeng Pan, Yanliang Zhou, Jidong Zhai, Xipeng Shen, Onur Mutlu, and Xiaoyong Du. 2021. G-TADOC: Enabling Efficient GPU-Based Text Analytics without Decompression. InICDE. 1679–1690

    2021

  62. [72]

    Feng Zhang, Jidong Zhai, Xipeng Shen, Onur Mutlu, and Wenguang Chen. 2018. Efficient document analytics on compressed data: method, challenges, algorithms, insights.Proc. VLDB Endow.11, 11 (July 2018), 1522–1535

    2018

  63. [73]

    Longtao Zhang, Ruoyu Li, Congrong Ren, Sheng Di, Jinyang Liu, Jiajun Huang, Robert Underwood, Pascal Grosset, Dingwen Tao, Xin Liang, Hanqi Guo, Franck Cappello, and Kai Zhao. 2025. LCP: Enhancing Scientific Data Management with Lossy Compression for Particles.Proc. ACM Manag....

    2025 doi

  64. [2022]

    In2022 IEEE 38th International Conference on Data Engineering (ICDE)

    Improving Prediction-Based Lossy Compression Dramatically via Ratio- Quality Modeling. In2022 IEEE 38th International Conference on Data Engineering (ICDE). 2494–2507. https://doi.org/10.1109/ICDE53745.2022.00232

    2022

Pith tools

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