Pith. sign in

REVIEW 3 major objections 6 minor 48 references

This paper claims that line density plots can show both dense trends and rare outliers by steering illumination with a per-line outlierness score.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 15:39 UTC pith:CW4JLVEP

load-bearing objection Solid methods paper for line-density shading; the outlierness metric has a real length-bias flaw and the perceptual gains are unproven, but it's worth refereeing. the 3 major comments →

arxiv 2512.16017 v2 pith:CW4JLVEP submitted 2025-12-17 cs.GR

Enhancing Line Density Plots with Outlier Control and Bin-based Illumination

Yumeng Xue , Bin Chen , Patrick Paetzold , Yunhai Wang , Christophe Hurter , Oliver Deussen This is my paper
classification cs.GR
keywords line density plotoutliernessbin-based similaritystructural normal mapluminance-only shadingCIEDE2000trajectory visualizationadaptive illumination
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Line density plots can hide the very things a reader might care about—individual paths and rare deviations—because dense corridors flood the image and binning breaks continuity. This paper argues that the fix is not to change the density map but to add a second visual channel: illumination that is driven by line structure rather than by density. The authors propose a per-line 'outlierness' score, built from an asymmetric bin-based similarity, and use it to decide which lines receive high-frequency shading detail. Two user sliders then let an analyst move emphasis between dominant trends and sparse outliers, and because shading touches only the luminance channel, the density colormap stays perceptually accurate. If this works as claimed, analysts gain an interactive tool to read both the forest and the rare trees in trajectory and time-series data.

Core claim

The central claim is that the perceived continuity and outlier visibility missing from line density plots can be restored in image space without corrupting the density encoding. The paper introduces a bin-based outlierness measure: each line is scored by how dissimilar its tangent orientations are to those of its neighbors, where similarity is an asymmetric integral of one line's density influence field along another's path. This score lets the method build a 'structural normal map' that layers high-frequency, trajectory-level normals over the usual density-gradient normals, with a prioritized replacement rule so every shaded pixel has a clear origin. Lighting is then made line-aware: per bi

What carries the argument

The load-bearing mechanism is the bin-based outlierness metric (Eq. 3-5): a non-commutative similarity between lines, computed as the normalized sum of one line's CDE influence field sampled along another line's pixels and weighted by directional alignment, then inverted into an average dissimilarity per line. This rank list drives a structural normal map (Eq. 10), which combines a density-gradient low-frequency normal map with a high-frequency map built from user-selected lines via prioritized replacement. Illumination uses per-bin light directions—set perpendicular to a dominant orientation obtained by weighted PCA (high-frequency pixels use the selected line's direction directly)—and appl

Load-bearing premise

The method's practical value rests on the assumption that the bin-based outlierness score ranks lines the way a human analyst would rank them; the paper supports this with the authors' visual inspection rather than a measured user study.

What would settle it

On a synthetic dataset with known ground truth—say, one dense corridor of parallel trajectories plus a few deliberately crossing lines—run the method with OutlierFocus=1, StructureEmphasis high, and check whether exactly the planted crossings are highlighted and the corridor remains unshaded. If the emphasis highlights unrelated geometry or fails to recover the planted outliers, the decoupling claim fails.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • An analyst can explore a trajectory dataset with two sliders—OutlierFocus and StructureEmphasis—to shift between flow-following and anomaly-highlighting views without recomputing the density plot.
  • Because hue and saturation are never altered, the density-to-color mapping stays interpretable, so the enhanced view can still be read as a density plot.
  • The outlierness metric is a standalone artefact: it can rank trajectories by shape congruence for queries like 'most typical path' or 'most anomalous path'.
  • The pipeline's linear precomputation and sub-second per-frame updates make interactive exploration feasible for medium-scale datasets (up to about 10,000 lines on modest hardware).
  • The luminance-only, direction-adaptive lighting recipe generalizes to other line-data visualizations, including scientific and medical line rendering.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Editorial: The paper validates the outlierness ranking only by the authors' visual inspection; a task-based user study would be needed to confirm that the metric's notion of 'outlier' matches human judgment under time pressure.
  • Editorial: The per-bin wPCA direction field is unsmoothed; on highly noisy data adjacent bins could get opposite light directions and produce salt-and-pepper flicker—the paper acknowledges this and suggests spatial regularization as future work.
  • Editorial: The CIEDE2000 budget gives a natural handle for automatic parameter selection: a system could clamp φ to keep distortion within a chosen tolerance while maximizing structural enhancement, turning the trade-off curve into a controller.
  • Editorial: Because the emphasis is relative to the current dataset (ranks normalize to [0,1]), a static image could exaggerate minor deviations in a clean dataset; interactive µ adjustment mitigates this, but static exports might need additional trend-vs-outlier encoding cues.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper presents a bin-based illumination method for line density plots. It introduces a bin-based outlierness metric (Eqs. 4–5) that ranks trajectories by local orientation congruence, uses this ranking to construct a structural normal map combining density-gradient normals with high-frequency line normals (Eqs. 7–10), computes per-bin light directions perpendicular to the dominant local orientation, and applies Lambertian shading only to the CIELAB L channel (Eqs. 11–12). The evaluation includes an ablation study, CIEDE2000 color-distortion measurements, runtime scaling, and case studies on vessel, stock, and temperature data.

Significance. The proposed pipeline is a plausible and useful extension of illumination-based density enhancement (e.g., VIDP) from point data to line data, with explicit user controls for balancing trend and outlier emphasis. The paper provides a concrete discretization of CDE, reproducible parameter settings, and quantitative measurements of color distortion and runtime. The CIEDE2000 analysis offers a practical parameter-selection guide. If the outlierness metric is made robust and the perceptual claims are backed by more than qualitative self-assessment, the method could become a valuable tool for trajectory and time-series visualization. The authors are candid about several limitations, which is a strength.

major comments (3)
  1. [Sec. 3.2, Eq. (4)] The normalization by |P_l'| makes sim(l,l') the average influence over the entire query line, including pixels far outside l's influence field where L_h≈0. Extending a line by a long tail in empty space therefore dilutes sim and inflates outlierness (Eq. 5) without changing local congruence. For example, a line with 100 px of overlap and 900 px of empty tail scores roughly 10× lower sim than the same 100 px overlap without the tail. Because the outlierness ranking drives the µ/σ emphasis (D2/D3), this length sensitivity can make the sliders highlight arbitrary long trajectories instead of meaningful deviations. Please normalize by the number of pixels with non-negligible influence, or otherwise make the measure invariant to irrelevant extension, and test ranking stability.
  2. [Sec. 4.1 and Sec. 5] The central claim that OutlierFocus/StructureEmphasis 'reveal' meaningful outliers is validated only by the authors' visual inspection (Sec. 4.1 states this explicitly) and Sec. 5 concedes that 'the absence of formal user studies limits claims regarding perceptual effectiveness.' The Chamfer/Hausdorff comparison in the supplementary material does not establish that the ranking matches analyst notions of outliers. I recommend adding a task-based evaluation with labeled outlier trajectories (e.g., the ALN and GLU-PRA cases in Sec. 4.4) or at least a quantitative ranking-quality measure against known ground truth.
  3. [Sec. 4.3 and Fig. 7] The abstract states 'supports interactive updates for up to 10,000 lines,' but Fig. 7 shows normal-map and lighting each below 1 s, and the text states 'the total update delay can approach 2 seconds.' For interactive exploration, 2 s per slider movement is at the boundary; this claim should be qualified. Additionally, the claimed O(n) outlierness computation is not substantiated: Eq. (4) contains an absolute value, so precomputed vector sums cannot give the exact sum of absolute dot products. Please provide the precise aggregation formula or report the empirical complexity honestly.
minor comments (6)
  1. [Sec. 3.2, Eq. (3)] The integral in Eq. (3) has a typographical artifact (a stray 'H'). Also, the notation l′ is reused in Eq. (6) for a normalized rank, which conflicts with its use as a line variable in Eqs. (3)–(4); please rename one of them.
  2. [Sec. 4.2, Fig. 6] The CIEDE2000 curves lack error bars and the Lambertian baseline is shown as a fixed value without specifying its φ or how it was chosen. Since this figure underpins the 'significantly lower color distortion' claim, please report variance across images and the exact baseline configuration.
  3. [Sec. 3.3] The definition of the spatial footprint B_i as an n×n pixel window is vague; specify how n relates to the discrete kernel size and whether the window is centered on the line pixels.
  4. [Sec. 3.5, Eq. (12)] The definitions of I_empty and I_min are not fully clear, and the role of negative φ (which the experiments use) deserves a brief explanation.
  5. [General] The paper repeatedly refers to 'supplementary material' (Secs. 1.1, 1.2, 3.2) but the submitted version does not include it; please ensure it is available to reviewers and readers.
  6. [Fig. 7] The y-axis label contains a placeholder ampersand: 'Time (ms) for & Normal Map Lighting' should be corrected.

Circularity Check

1 steps flagged

No significant circularity: the outlierness-based shading pipeline is self-contained; the only by-construction claim is the CIEDE2000 color-fidelity advantage, which the paper itself labels 'by design'.

specific steps
  1. other [Sec. 4.2 (Color Distortions), Fig. 6]
    "This significant difference is by design. The baseline’s high distortion (see dashed lines) is a direct consequence of applying Lambertian shading—which acts as a form of brightness scaling—directly in the RGB color space. This single operation non-linearly alters all three dimensions of the perceptual CIELAB space (L, a, and b), causing visible shifts in both hue and luminance. Our method, by contrast, is designed to apply illumination only to the L channel, which results in fundamentally lower color distortion."

    The headline claim of lower CIEDE2000 distortion is not an independent empirical prediction: since the method constrains CIELAB a/b channels to remain unchanged while the Lambertian baseline scales all RGB channels, lower ΔE00 follows by construction from the design. The paper explicitly says the difference is 'by design.' This is a consistency check of the design constraint rather than a derived result, and it does not infect the structural normal-map or outlier-emphasis contribution.

full rationale

The core derivation chain—bin-based outlierness (Eqs. 3-5), ranked selection (Eq. 6), structural normal map (Eqs. 7-10), and local lighting (Eq. 11)—is not circular: no parameter is fitted to the output it is said to predict, and no uniqueness theorem or hidden ansatz is imported from the authors' own prior work. The self-citations to Chen et al. [8] for luminance-only shading and Xue et al. [42] for colorization are to published prior components and are not load-bearing for the central claim; the paper also provides its own rationale for the luminance-channel choice. The color-fidelity comparison flagged above is the one place where an evaluation claim reduces to a design constraint, but the paper explicitly acknowledges this. The remaining weaknesses are validation gaps, not circularity: Sec. 4.1 states the ablation comparisons are 'based on qualitative visual assessment by the authors, not on formal perceptual metrics or a controlled user study,' and Sec. 5 concedes 'the absence of formal user studies limits claims regarding perceptual effectiveness.' The outlierness metric's sensitivity to line length (Eq. 4's normalization) is a robustness concern, not a circular reduction. Overall the method is self-contained and mostly honestly evaluated, so a low score is appropriate.

Axiom & Free-Parameter Ledger

6 free parameters · 7 axioms · 3 invented entities

The pipeline introduces no fitted constants; all four interaction parameters are user knobs with example values. The outlierness metric and structural normal map are algorithmic constructs, not physical postulates. The main unvalidated assumptions are perceptual: that the metric equals visual anomaly salience and that luminance-only shading preserves readability as well as color fidelity.

free parameters (6)
  • OutlierFocus µ = user-controlled; examples 0.0–1.0 (0.6, 1.0, 0.25)
    Selects which outlierness rank is emphasized; results in Figs 1, 8, 9 depend on it.
  • StructureEmphasis σ = user-controlled; examples 0.05, 0.4, 0.5
    Fraction of top-ranked lines contributing high-frequency normals; controls detail level.
  • Normal scaling η = 3.0 (vessel), 1.0 (stock/temperature)
    In Eq. (7) scales the z-component of low-frequency normals; exaggerates slope.
  • Illumination strength φ = -20 or -25
    In Eq. (12) controls amount of luminance shading; trade-off with color distortion.
  • CDE kernel size / bandwidth h = 3×3 or 5×5 pixels
    Controls density smoothness and line thickness; chosen per dataset.
  • Influence diffusion range = 5 pixels on either side
    Range of distance field for line similarity; a computational approximation.
axioms (7)
  • standard math Lampe & Hauser CDE field L_h is a valid continuous influence field for line similarity
    Eqs. (1)-(2), (3)-(4) build on accepted density estimation from curve density estimates.
  • domain assumption Gradient of the density field gives a meaningful low-frequency normal map
    Eq. (7) inherits the density-gradient normal model from prior shading methods [33,34].
  • domain assumption Lambertian shading with a 60° elevation angle is perceptually appropriate
    Sec. 3.4 cites O'Shea et al. [27]; fixed elevation but not validated specifically for line plots.
  • domain assumption Luminance-only shading in CIELAB preserves the density colormap better without harming structural readability
    Sec. 3.5 follows VIDP [8]; color distortion is measured, but readability is not user-tested.
  • domain assumption The outlierness score correctly ranks visual distinctiveness of trajectories
    Eq. (5) definition; authors compare visually to Chamfer/Hausdorff but provide no user study or ground truth.
  • standard math Weighted PCA on local tangent vectors gives a stable dominant orientation
    Sec. 3.4 uses closed-form PCA on local tangents to set light direction; the authors note possible noise artifacts.
  • domain assumption Random subsampling of trajectories preserves the visual structure of the dataset
    Case studies use 3,000 or 2,000 sampled lines from larger datasets; no sensitivity analysis of sampling is provided.
invented entities (3)
  • Bin-based line outlierness score no independent evidence
    purpose: Ranks each trajectory by average dissimilarity to its neighbors; drives user-controlled emphasis.
    Algorithmically defined in Eqs. (3)-(5), but no external perceptual validation; depends on chosen influence field and neighbor criterion.
  • Structural normal map n_structure no independent evidence
    purpose: Combines low-frequency density normals with high-frequency line normals; decouples shading from density.
    Eq. (10) composes n_low and n_high; a rendering construct with no independent physical referent.
  • Per-bin light direction no independent evidence
    purpose: Local Lambertian light perpendicular to dominant line orientation, removing fixed-light orientation bias.
    A rendering choice; no external evidence that this improves perception beyond the authors' qualitative comparisons.

pith-pipeline@v1.3.0-alltime-deepseek · 18578 in / 13298 out tokens · 129115 ms · 2026-08-03T15:39:49.001005+00:00 · methodology

0 comments
read the original abstract

Density plots effectively summarize large numbers of points, which would otherwise lead to severe overplotting in, for example, a scatter plot. However, when applied to line-based datasets, such as trajectories or time series, density plots alone are insufficient, as they disrupt path continuity, obscuring smooth trends and rare anomalies. We propose a bin-based illumination model that decouples structure from density to enhance flow and reveal sparse outliers while preserving the original colormap. We introduce a bin-based outlierness metric to rank trajectories. Guided by this ranking, we construct a structural normal map and apply locally-adaptive lighting in the luminance channel to highlight chosen patterns -- from dominant trends to atypical paths -- with acceptable color distortion. Our interactive method enables analysts to prioritize main trends, focus on outliers, or strike a balance between the two. We demonstrate our method on several real-world datasets, showing it reveals details missed by simpler alternatives, achieves significantly lower CIEDE2000 color distortion than standard shading, and supports interactive updates for up to 10,000 lines.

Figures

Figures reproduced from arXiv: 2512.16017 by Bin Chen, Christophe Hurter, Oliver Deussen, Patrick Paetzold, Yumeng Xue, Yunhai Wang.

Figure 1
Figure 1. Figure 1: Comparison of line-density visualization methods on vessel trajectories. (a) Plain density plot highlights only high-density [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Pipeline for enhanced, discretized line density plots: (a) Initial density plot; (b) Normal maps, including low- and high-frequency normal [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The sensitivity of traditional Lambertian shading to a fixed global [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Illustration of bin-based line similarity: (a) Construction of the [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Ablation and comparative study on the vessel dataset [ [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Color difference (∆E00) as a function of the illumination parameter φ. Solid lines trace our method’s distortion per dataset, while dashed lines show the higher, fixed distortion of the Lambertian baseline. The gray line marks our acceptable tolerance threshold of ∆E00 = 3.0, based on printing industry benchmarks [19]. key consideration is the trade-off between the strength of this enhance￾ment and the res… view at source ↗
Figure 7
Figure 7. Figure 7: Scalability of computation time vs. number of input lines for [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Our method with different parameters to highlight different degrees of outlierness for the Hellenic Trench AIS dataset [ [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Comparison of our method using a multi-hue colormap on the stock dataset [ [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Application of our method on the temperature dataset [ [PITH_FULL_IMAGE:figures/full_fig_p008_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Our method allows us to highlight differences between clusters. Here, we show the vessel trajectories with five clusters and hue-preserved [PITH_FULL_IMAGE:figures/full_fig_p008_11.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

48 extracted references · 7 canonical work pages

  1. [1]

    https://www.nyse.com/market-data/ historical

    New york stock exchange. https://www.nyse.com/market-data/ historical. Accessed: 2025-09-01. 7, 8

    2025

  2. [2]

    http://www.rcc-acis.org/docs_webservices

    Acis web services. http://www.rcc-acis.org/docs_webservices. html, 2017. Accessed: 2025-09-01. 8, 9

    2017

  3. [3]

    https://ca.finance.yahoo

    Yahoo finance: Glu-pra stock price. https://ca.finance.yahoo. com/quote/GLU-PRA/, 2025. Accessed: 2025-09-01. 9

    2025

  4. [4]

    Akmal Butt and P

    M. Akmal Butt and P. Maragos. Optimum design of chamfer distance transforms.IEEE Transactions on Image Processing, 7(10):1477–1484,

  5. [5]

    Arthur Van, F

    G. Arthur Van, F. Staals, M. Löffler, J. Dykes, and B. Speckmann. Multi- granular trend detection for time-series analysis.IEEE Transactions on Visualization and Computer Graphics, 23(1):661–670, 2017. doi: 10. 1109/TVCG.2016.2598619 2

    arXiv 2017

  6. [6]

    H. Bao, X. Chen, K. Lu, C.-W. Fu, J.-D. Fekete, and Y . Wang. Bi-scale density-plot enhancement based on variance-aware filter.Computers & Graphics, 127:104180, 2025. doi: 10.1016/j.cag.2025.104180 2

    arXiv 2025

  7. [7]

    D. B. Carr, R. J. Littlefield, W. L. Nicholson, and J. S. Littlefield. Scat- terplot matrix techniques for large n.Journal of the American Statisti- cal Association, 82(398):424–436, 1987. doi: 10.1080/01621459.1987. 10478445 2

    arXiv 1987

  8. [8]

    X. Chen, Y . Wang, H. Bao, K. Lu, J. Jo, C.-W. Fu, and J.-D. Fekete. Visualization-driven illumination for density plots.IEEE Transactions on Visualization and Computer Graphics, 31(2):1631–1644, 2025. doi: 10. 1109/TVCG.2024.3495695 2, 3, 4, 5

    arXiv 2025

  9. [9]

    Eichelbaum, M

    S. Eichelbaum, M. Hlawitschka, and G. Scheuermann. Lineao—improved three-dimensional line rendering.IEEE Transactions on Visualization and Computer Graphics, 19(3):433–445, 2013. doi: 10.1109/TVCG.2012.142 9

    work page doi:10.1109/tvcg.2012.142 2013

  10. [10]

    W. D. Ellis.A Source Book Of Gestalt Psychology (1st ed.). Routledge,

  11. [11]

    D. Feng, L. Kwock, Y . Lee, and R. Taylor. Matching visual saliency to confidence in plots of uncertain data.IEEE Transactions on Visualization and Computer Graphics, 16(6):980–989, 2010. doi: 10.1109/TVCG.2010. 176 2

    doi:10.1109/tvcg.2010 2010

  12. [12]

    Frantzis, R

    A. Frantzis, R. Leaper, P. Alexiadou, A. Prospathopoulos, and D. Lekkas. Hellenic trench ais data. 2018. doi: 10.17882/57040 6, 7, 8

    doi:10.17882/57040 2018

  13. [13]

    Frantzis, R

    A. Frantzis, R. Leaper, P. Alexiadou, A. Prospathopoulos, and D. Lekkas. Shipping routes through core habitat of endangered sperm whales along the hellenic trench, greece: Can we reduce collision risks?PLOS ONE, 14(2):1–21, 2019. doi: 10.1371/journal.pone.0212016 7

    work page doi:10.1371/journal.pone.0212016 2019

  14. [14]

    Hochheiser and B

    H. Hochheiser and B. Shneiderman. Dynamic query tools for time se- ries data sets: Timebox widgets for interactive exploration.Information Visualization, 3(1):1–18, 2004. doi: 10.1057/palgrave.ivs.9500061 3

    doi:10.1057/palgrave.ivs.9500061 2004

  15. [15]

    Hurter, B

    C. Hurter, B. Tissoires, and S. Conversy. Fromdady: Spreading aircraft trajectories across views to support iterative queries.IEEE Transactions on Visualization and Computer Graphics, 15(6):1017–1024, 2009. doi: 10 .1109/TVCG.2009.145 2, 3

    2009

  16. [16]

    Huttenlocher, G

    D. Huttenlocher, G. Klanderman, and W. Rucklidge. Comparing images using the hausdorff distance.IEEE Transactions on Pattern Analysis and Machine Intelligence, 15(9):850–863, 1993. doi: 10.1109/34.232073 4

    doi:10.1109/34.232073 1993

  17. [17]

    D. F. Jerding and J. T. Stasko. The information mural: a technique for displaying and navigating large information spaces.IEEE Transactions on Visualization and Computer Graphics, 4(3):257–271, 1998. doi: 10. 1109/2945.722299 2

    arXiv 1998

  18. [18]

    O. D. Lampe and H. Hauser. Curve Density Estimates.Computer Graphics Forum, 30(3):633–642, 2011. doi: 10.1111/j.1467-8659.2011.01912.x 2, 4

    arXiv 2011

  19. [19]

    H. X. Liu, B. Wu, Y . Liu, M. Huang, and Y . F. Xu. A discussion on printing color difference tolerance by ciede2000 color difference formula. InAdvances in Printing and Packaging Technologies (Applied Mechanics and Materials), vol. 262, pp. 96–99. Trans Tech Publications Ltd, 2 2013. doi: 10.4028/www.scientific.net/AMM.262.96 6

    work page doi:10.4028/www.scientific.net/amm.262.96 2013

  20. [20]

    Mannino and A

    M. Mannino and A. Abouzied. Expressive time series querying with hand- drawn scale-free sketches. InProceedings of the 2018 CHI Conference on Human Factors in Computing Systems, CHI ’18, 13 pages, p. 1–13. Association for Computing Machinery, New York, NY , USA, 2018. doi: 10.1145/3173574.3173962 3

    arXiv 2018

  21. [21]

    Matejka, F

    J. Matejka, F. Anderson, and G. Fitzmaurice. Dynamic opacity optimiza- tion for scatter plots. InProceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, CHI ’15, 4 pages, p. 2707–2710. ACM, New York, 2015. doi: 10.1145/2702123.2702585 2

    arXiv 2015

  22. [22]

    Mayorga and M

    A. Mayorga and M. Gleicher. Splatterplots: Overcoming overdraw in scatter plots.IEEE Transactions on Visualization and Computer Graphics, 19(9):1526–1538, 2013. doi: 10.1109/TVCG.2013.65 2, 4

    doi:10.1109/tvcg.2013.65 2013

  23. [23]

    Micallef, G

    L. Micallef, G. Palmas, A. Oulasvirta, and T. Weinkauf. Towards percep- tual optimization of the visual design of scatterplots.IEEE Transactions on Visualization and Computer Graphics, 23(6):1588–1599, 2017. doi: 10 .1109/TVCG.2017.2674978 2

    arXiv 2017

  24. [24]

    Mokrzycki and M

    W. Mokrzycki and M. Tatol. Color difference delta e - a survey.Machine Graphics and Vision, 20:383–411, 04 2011. 6

    2011

  25. [25]

    Moritz and D

    D. Moritz and D. Fisher. Visualizing a million time series with the density line chart, 2018. doi: 10.48550/arXiv.1808.06019 2

    Pith/arXiv arXiv doi:10.48550/arxiv.1808.06019 2018

  26. [26]

    Novotny and H

    M. Novotny and H. Hauser. Outlier-preserving focus+context visualization in parallel coordinates.IEEE Transactions on Visualization and Computer Graphics, 12(5):893–900, 2006. doi: 10.1109/TVCG.2006.170 2

    work page doi:10.1109/tvcg.2006.170 2006

  27. [27]

    J. P. O’Shea, M. S. Banks, and M. Agrawala. The assumed light direction for perceiving shape from shading. InProceedings of the 5th Symposium on Applied Perception in Graphics and Visualization, APGV ’08, 8 pages, p. 135–142. Association for Computing Machinery, New York, NY , USA,

  28. [28]

    B. T. Phong.Illumination for computer generated pictures, p. 95–101. Association for Computing Machinery, New York, NY , USA, 1998. doi: 10.1145/280811.280980 1, 4, 5

    arXiv 1998

  29. [29]

    J. F. Pinto da Costa, H. Alonso, and L. Roque. A weighted principal component analysis and its application to gene expression data.IEEE/ACM Transactions on Computational Biology and Bioinformatics, 8(1):246–252,

  30. [30]

    Historical stock data for former amer- ican lorain corporation (ticker: Aln)

    Planet Green Holdings Corp. Historical stock data for former amer- ican lorain corporation (ticker: Aln). Information accessed through historical financial data providers. Formerly traded as ALN on NYSE American., 2025. The company American Lorain Corporation (for- merly ALN) was renamed to Planet Green Holdings Corp. (ticker: PLAG) in September 2018. The...

    2025

  31. [31]

    Pomerenke, F

    D. Pomerenke, F. L. Dennig, D. A. Keim, J. Fuchs, and M. Blumenschein. Slope-dependent rendering of parallel coordinates to reduce density distor- tion and ghost clusters. In2019 IEEE Visualization Conference (VIS), pp. 86–90, 2019. doi: 10.1109/VISUAL.2019.8933706 2

    arXiv 2019

  32. [32]

    Rusinkiewicz, M

    S. Rusinkiewicz, M. Burns, and D. DeCarlo. Exaggerated shading for depicting shape and detail. InACM SIGGRAPH 2006 Papers, SIGGRAPH ’06, 7 pages, p. 1199–1205. Association for Computing Machinery, New York, NY , USA, 2006. doi: 10.1145/1179352.1142015 5

    arXiv 2006

  33. [33]

    Scheepens, N

    R. Scheepens, N. Willems, H. van de Wetering, G. Andrienko, N. An- drienko, and J. J. van Wijk. Composite density maps for multivariate trajectories.IEEE Transactions on Visualization and Computer Graphics, 17(12):2518–2527, 2011. doi: 10.1109/TVCG.2011.181 1, 2, 3, 4

    doi:10.1109/tvcg.2011.181 2011

  34. [34]

    Scheepens, N

    R. Scheepens, N. Willems, H. van de Wetering, and J. J. van Wijk. Inter- active visualization of multivariate trajectory data with density maps. In 2011 IEEE Pacific Visualization Symposium, pp. 147–154, 2011. doi: 10. 1109/PACIFICVIS.2011.5742384 1, 2, 4

    arXiv 2011

  35. [35]

    Sharma, W

    G. Sharma, W. Wu, and E. N. Dalal. The ciede2000 color-difference formula: Implementation notes, supplementary test data, and mathematical observations.Color Research & Application, 30(1):21–30, 2005. doi: 10. 1002/col.20070 6

    2005

  36. [36]

    B. W. Silverman.Density estimation for statistics and data analysis, vol. 26. CRC press, 1986. 2

    1986

  37. [37]

    Thöny and R

    M. Thöny and R. Pajarola. Vector map constrained path bundling in 3d environments. InProceedings of the 6th ACM SIGSPATIAL International Workshop on GeoStreaming, IWGS ’15, 10 pages, pp. 33–42. ACM, New York, 2015. doi: 10.1145/2833165.2833168 2

    arXiv 2015

  38. [38]

    Trautner, F

    T. Trautner, F. Bolte, S. Stoppel, and S. Bruckner. Sunspot Plots: Model- based Structure Enhancement for Dense Scatter Plots.Computer Graphics Forum, 2020. doi: 10.1111/cgf.14001 3

    doi:10.1111/cgf.14001 2020

  39. [39]

    Trautner, M

    T. Trautner, M. Sbardellati, S. Stoppel, and S. Bruckner. Honeycomb Plots: Visual Enhancements for Hexagonal Maps. InVision, Modeling, and Visualization. The Eurographics Association, 2022. doi: 10.2312/vmv .20221205 3 10 © 2025 IEEE. This is the author’s version of the article that has been published in IEEE Transactions on Visualization and Computer Gra...

    work page doi:10.2312/vmv 2022

  40. [40]

    H. Wickham. Bin-summarise-smooth: a framework for visualising large data.had. co. nz, Tech. Rep, 2013. 2

    2013

  41. [41]

    Willems, H

    N. Willems, H. Van De Wetering, and J. J. Van Wijk. Visualization of vessel movements.Computer Graphics Forum, 28(3):959–966, 2009. doi: 10.1111/j.1467-8659.2009.01440.x 1, 2, 3

    arXiv 2009

  42. [42]

    Y . Xue, P. Paetzold, R. Kehlbeck, B. Chen, K. C. Kwan, Y . Wang, and O. Deussen. Reducing ambiguities in line-based density plots by image- space colorization.IEEE Transactions on Visualization and Computer Graphics, 30(1):825–835, 2024. doi: 10.1109/TVCG.2023.3327149 1, 2, 5, 9

    arXiv 2024

  43. [43]

    W. Zeng, Q. Shen, Y . Jiang, and A. Telea. Route-aware edge bundling for visualizing origin-destination trails in urban traffic.Computer Graphics Forum, 38(3):581–593, 2019. doi: 10.1111/cgf.13712 2

    doi:10.1111/cgf.13712 2019

  44. [44]

    Y . Zhao, Y . Wang, J. Zhang, C.-W. Fu, M. Xu, and D. Moritz. Kd-box: Line-segment-based kd-tree for interactive exploration of large-scale time- series data.IEEE Transactions on Visualization and Computer Graphics, 28(1):890–900, 2022. doi: 10.1109/TVCG.2021.3114865 2, 3 11

    arXiv 2022

  45. [1998]

    doi: 10.1109/83.718487 4

    doi:10.1109/83.718487

  46. [2008]

    doi: 10.1145/1394281.1394306 4, 5

    arXiv

  47. [2009]

    doi: 10.1109/TCBB.2009.61 5

    work page doi:10.1109/tcbb.2009.61 2009

  48. [2013]

    doi: 10.4324/9781315009247 1

    work page doi:10.4324/9781315009247