Pith. sign in

REVIEW 5 major objections 5 minor 79 references

Transforming CCTV cameras into NO$_2$ sensors at city scale for adaptive policymaking

T0 review · 5 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read CCTV cameras can act as citywide NO2 sensors

desk verdict Good idea, unverified claim: no ablation shows the CCTV features beat weather-plus-interpolation, validation is in-sample, and the data counts disagree — but the pipeline deserves a serious referee. read the letter →

arxiv 2501.00056 v1 pith:PZW435S3 submitted 2024-12-28 cs.LG cs.AIcs.CY

Mohamed R. Ibrahim , Terry Lyons This is my paper · ORCID
classification cs.LGcs.AIcs.CY
keywords airqualitymonitoringNO2predictionCCTVtrafficanalysisgraphneuralnetworkspathsignaturecity-scalesensingurbanpolicyLondon
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 existing CCTV cameras, originally installed for traffic monitoring, can be repurposed as pseudo-NO2 sensors. It uses 133 million frames from 907 London cameras to extract road-user flows, summarizes each hourly stream with a path signature, and feeds the result through a graph neural network to predict hourly NO2 concentrations across the city, including at locations without any pollution sensor. If this works, cities with dense camera networks but sparse air-quality monitors could get street-level pollution maps without buying new hardware. The paper also claims that different traffic modes affect NO2 on different time lags, up to six hours, and argues that this undercuts policies that simply push truck traffic into the night.

What carries the argument

The load-bearing object is the path signature, a summary of a multidimensional stream that is invariant to reparameterization; here it is computed to depth N=3 from a 13-channel tensor holding modal flows and stationary statuses across 11 file increments per hour. The signature compresses the order of road-user events in a camera scene into a fixed-size feature while retaining the information needed for prediction. That feature is combined with environmental and spatial variables and passed through a graph-to-graph neural network with two graphs: a K-nearest-neighbour traffic graph (K=10) for the cameras, chosen to mimic the road network, and a fully connected graph for the NO2 sensors, chosen because air can diffuse freely. The model learns to generate NO2 values on the sensor graph from signals on the camera graph, which is what makes prediction at unsensed locations possible.

What would settle it

Hold out a random subset of NO2 sensors from training and run portable reference monitors at those locations and at several sensor-free camera sites; if the graph model's predictions at those points are no better than a spatial interpolation baseline, the CCTV signal is not carrying the prediction. A simpler lab check is to retrain the model with the traffic-flow channels replaced by shuffled noise and compare validation R2: if it barely drops, the camera data contributes little.

Watch

Extended reading notes

Core claim

The central claim is that ground-level NO2 can be inferred from camera-derived traffic dynamics through a graph-to-graph mapping: input features at 907 camera nodes (per-mode traffic counts, stationary statuses, congestion, plus weather and locational factors) are transformed into an NO2 field at 144 sensor locations, with no requirement that camera and sensor positions coincide. The strongest reported city-wide model, a Conv1D multi-branch graph network with signature inputs of depth three, reaches an MSLE of 0.0285, an MAE of 0.5197, and an R2 of 0.82 on validation data. The authors interpret the learned temporal lags, such as stationary trucks showing a one-hour Granger-causal effect and moving trucks a five-hour effect, as evidence that NO2 dynamics are tied to the chemical lifetime of NO2 and to the timing of traffic. The practical conclusion is that a hardware-free, CCTV-based sensing layer is feasible for urban NO2 monitoring and can inform adaptive policy.

Load-bearing premise

The load-bearing premise is that the K-nearest-neighbour traffic graph and the fully connected sensor graph capture the real physics of NO2 dispersion, and that traffic features at 907 cameras plus weather data are enough to determine NO2 at 144 sensor locations and everywhere in between.

Editorial extensions

If this is right

  • Cities with extensive CCTV coverage but few air-quality monitors could generate hourly NO2 maps with no new sensing hardware.
  • The reported Granger lags imply that moving heavy truck traffic to night hours will not necessarily reduce morning peak exposure; policy should weigh total emissions across all hours.
  • Because camera and sensor locations do not have to align, the graph-to-graph approach can in principle be transferred to other cities and to other pollutants.
  • With inference under 0.1 seconds per hour, the prediction side is fast enough for near-real-time monitoring if the video-processing front end can be scaled.

Reading between the lines

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

  • The validation is performed only at the 144 sensor locations, so a decisive field test would place portable NO2 monitors in camera-rich, sensor-poor blocks and compare model output against them; if accuracy collapses, the model is mostly interpolating sensor readings.
  • Because the NO2 graph is fully connected, the network could in principle learn a purely spatial smooth of the sensor targets; shuffling or zeroing the traffic features while retraining would reveal how much of the prediction actually comes from the CCTV signal.
  • The same architecture might transfer to PM2.5, but its different atmospheric lifetime means the lag structure and graph assumptions would need to be re-estimated rather than reused.
  • The policy criticism based on Granger causality is predictive rather than causal; an intervention study, such as a real change in lorry routing times, would be needed to confirm that nighttime truck bans worsen morning NO2.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes a pipeline that extracts traffic modal flows from London CCTV footage, summarizes hourly streams with path signatures, and uses graph neural networks to predict hourly NO2 concentrations at 144 sensor locations. The authors claim that CCTV cameras act as pseudo-NO2 sensors, enabling city-wide pollutant maps and policy insights such as temporal lags of traffic effects of up to 6 hours. The empirical section reports validation metrics for 20 model configurations, along with spatial regression and Granger causality analyses.

Significance. The problem is important, and assembling a large CCTV dataset with per-mode traffic counts is non-trivial. The combination of path signatures and graph-to-graph learning is a reasonable technical approach. However, the central claim is not established by the reported experiments. No ablation removes the CCTV-derived features, no independent test set or temporal/spatial holdout is used, and the reported validation metrics are therefore not sufficient to separate the contribution of camera data from spatial interpolation and environmental covariates. The paper also contains internal inconsistencies in data counts and in the Granger lag results. If the missing experiments were performed and the claims held, the work would be a useful contribution to urban air-quality monitoring.

major comments (5)
  1. [Section 2.6, Table S5] The paper never tests a model that excludes the CCTV traffic inputs: all 20 models in Table S5 include traffic flow features, with or without signatures. Because NO2 is strongly spatially autocorrelated and the model also receives weather, land use, and a fully connected NO2 sensor graph, the reported validation performance may be achievable without any camera data. The authors must add a baseline that removes or permutes the CCTV-derived features while keeping everything else identical, and report the resulting MSLE and R2. This is necessary to support the paper's core claim that CCTV cameras act as pseudo-NO2 sensors.
  2. [Section 4, Table S5] All metrics are reported on the validation set, and the models are selected by early stopping on validation loss. There is no description of how the validation split is formed (random, temporal, or spatial), no independent test set, and no held-out period or held-out sensor locations. A random split that mixes hours from the same sensors in training and validation can yield optimistic R2 values due to spatial and temporal autocorrelation. The authors should report results on a held-out temporal period (for example, a set of days) and on a held-out spatial subset of NO2 sensors to support city-wide generalization claims.
  3. [Section 2.1 vs Methods] Data counts are contradictory. Section 2.1 states 133,132,866 frames from 907 locations over 112 hours; the Methods (Materials) state 65,493,858 frames from 892 locations over 56 hours, and Section 2.2 says n=906. The abstract repeats the '133 million frames' figure. The paper must reconcile these numbers, as the dataset description is the basis for the headline claim and for reproducing the experiments.
  4. [Section 2.4, Table 3] The abstract and text claim Granger-causal lags 'of up to 6 hours', but Table 3 reports only lags 1 through 5, with no lag-6 column. Many reported lags are significant in some tests but not others (for example, Car count at Hackney 1.01301002, Lag 5: alpha p=0.1267 and delta p=0.1267, but beta p<0.0001). The analysis is restricted to a handful of camera locations, yet the policy implications in Section 3 are city-wide. Either report the 6-hour lag results or correct the claim, and discuss the inconsistency of significance across the four tests.
  5. [Section 2.6] The choice of a fully connected NO2 graph and a KNN (K=10) traffic graph is asserted with a physical rationale ('air can diffuse freely') but not empirically validated. In particular, the fully connected target graph may enable the model to interpolate NO2 directly from neighboring sensors, making the traffic input redundant. The authors should test alternative graph constructions (for example, distance-based or thresholded graphs for the NO2 sensors) and compare performance, ideally as part of the ablation described in the first major comment.
minor comments (5)
  1. [Section 2.2] The camera count is given as n=906 in Section 2.2, but Section 2.1 and the abstract refer to 907 cameras; please correct the inconsistency.
  2. [Figure 5] Panels B and D have identical captions ('scatter plot for the actual and predicted data for all sensor locations and all dates'); clarify which panel corresponds to training and which to validation, or indicate the actual distinction.
  3. [Table 1] Several p-values are reported as 0.0000 (for example, 'car count' and 'rainfall'); these should be reported as p<0.0001 or with a more precise notation.
  4. [Equation (16)] The definition of the truncated signature is garbled; the notation 'NM' and 'hat S' is not standard, and the expression for S^n is not an iterated integral. Please provide a standard reference or a correct formulation.
  5. [Table S5 vs Section 2.6] The labels 'Task 1' and 'Task 2' in Table S5 are reversed relative to the descriptions in Section 2.6 (Methods), where Task 1 is the NO2 surface and Task 2 is a single location. Please align the terminology.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the NO2 predictions are supervised model outputs, not re-statements of the training labels or of the inputs by construction.

full rationale

The paper's derivation chain is: CCTV frames -> object detection/tracking -> traffic modal flows -> path-signature features -> graph neural network trained with NO2 sensor labels -> predicted NO2 at sensor locations -> triangulated surface. The NO2 ground truth enters only as a supervised training target; it is not an input feature of the prediction function at inference, and no equation defines the predicted NO2 in terms of observed NO2. The self-citations (TopView for bird's-eye-view projection and rough-path signature references) are implemented with explicit equations (homography and signature transforms) and are not invoked as a forced uniqueness argument or as the sole justification for the central CCTV-to-NO2 claim. The absence of a traffic-ablation or a spatially/temporally held-out benchmark is a real correctness and generalizability risk, but it is not circularity by construction: the reported validation metrics are out-of-sample predictions on a validation split, not fitted parameters renamed as predictions. The reader's concern that the NO2 surface is 'a fitted function of the inputs' applies to every supervised regression and does not constitute a definitional reduction. The limitations section openly notes placement bias and sensor coverage assumptions, which further confirms the authors are not claiming a first-principles derivation. Therefore no specific circular step can be quoted, and the appropriate score is 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests entirely on a fitted graph neural network (GNN) plus statistical lag analyses; there are no physical derivations. The free parameters are the model weights, the KNN neighborhood size, and the signature depth. The axioms are domain assumptions about representativeness of CCTV data, accuracy of NO2 sensors, adequacy of the graph structures, and sufficiency of signature features. No new physical entities are introduced.

free parameters (3)
  • Graph neural network weights = Not disclosed (millions of parameters; e.g., 120M for attention model)
    The predictive model's weights are trained on the collected NO2 labels; predictions are outputs of these fitted weights, so the 'prediction' is a fit, not an independent derivation.
  • K in K-nearest-neighbour spatial weight = 10 for traffic graph; 50 used in best task-2 model; 10 for best task-1 model
    Section 4 (Spatial weight) states K=10 was 'the best approximation'; Table 5 shows K=50 for the best city-wide model, indicating K is tuned rather than fixed.
  • Signature depth N = 3
    The path signature is truncated at depth 3 (Section 4, 'Signature of paths') without an ablation of higher depths; a modeling choice that affects predictive features.
assumptions (4)
  • domain assumption CCTV-observed traffic flows at 907 locations are representative of city-wide traffic emissions.
    The entire method treats camera counts as the spatial basis for NO2 inference, though Section 3.1 admits cameras are typically in high-traffic areas and may bias predictions.
  • domain assumption Hourly NO2 sensor measurements are accurate ground truth at their locations.
    The models are trained and validated against these measurements (Section 4, 'Hourly NO2 data'), with no discussion of sensor calibration error.
  • ad hoc to paper The KNN (K=10) and fully-connected graphs represent physical NO2 dispersion.
    Section 2.6 asserts the sensor graph is fully connected 'because air can diffuse freely', a simplification adopted without comparison to meteorological transport models.
  • ad hoc to paper Path signature at depth 3 preserves the information in the traffic stream relevant to NO2.
    Signatures are used as input features (Section 4, 'Signature of paths'); no theorem links signature depth 3 to NO2 chemistry, and Table 5 shows a no-signature model performs similarly at Task 1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transforming CCTV cameras into NO$_2$ sensors at city scale for adaptive policymaking." pith.science (2026). https://pith.science/paper/PZW435S3

@misc{pith2026250100056,
  author       = {Pith},
  title        = {Pith review of: Transforming CCTV cameras into NO$_2$ sensors at city scale for adaptive policymaking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PZW435S3}},
  note         = {Machine review of arXiv:2501.00056}
}
read the original abstract

Air pollution in cities, especially NO\textsubscript{2}, is linked to numerous health problems, ranging from mortality to mental health challenges and attention deficits in children. While cities globally have initiated policies to curtail emissions, real-time monitoring remains challenging due to limited environmental sensors and their inconsistent distribution. This gap hinders the creation of adaptive urban policies that respond to the sequence of events and daily activities affecting pollution in cities. Here, we demonstrate how city CCTV cameras can act as a pseudo-NO\textsubscript{2} sensors. Using a predictive graph deep model, we utilised traffic flow from London's cameras in addition to environmental and spatial factors, generating NO\textsubscript{2} predictions from over 133 million frames. Our analysis of London's mobility patterns unveiled critical spatiotemporal connections, showing how specific traffic patterns affect NO\textsubscript{2} levels, sometimes with temporal lags of up to 6 hours. For instance, if trucks only drive at night, their effects on NO\textsubscript{2} levels are most likely to be seen in the morning when people commute. These findings cast doubt on the efficacy of some of the urban policies currently being implemented to reduce pollution. By leveraging existing camera infrastructure and our introduced methods, city planners and policymakers could cost-effectively monitor and mitigate the impact of NO\textsubscript{2} and other pollutants.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

79 extracted references · 76 canonical work pages

  1. [1]

    Nat Commun 11, 5366 (2020)

    Sun, L., Chen, J., Li, Q., Huang, D.: Dramatic uneven urbanization of large cities throughout the world in recent decades. Nat Commun 11, 5366 (2020)

    work page 2020

  2. [2]

    Nature 604, 104–110 (2022)

    Coutrot, A., et al.: Entropy of city street networks linked to future spatial navigation ability. Nature 604, 104–110 (2022)

    work page 2022

  3. [3]

    Nat Commun 13, 7977 (2022)

    Anza-Ramirez, C., et al.: The urban built environment and adult bmi, obesity, and diabetes in latin american cities. Nat Commun 13, 7977 (2022)

    work page 2022

  4. [4]

    Social Science & Medicine 75, 1469–1476 (2012)

    Badland, H.M., et al.: Association of neighbourhood residence and preferences with the built environment, work-related travel behaviours, and health impli- cations for employed adults: Findings from the urban study. Social Science & Medicine 75, 1469–1476 (2012)

    work page 2012

  5. [5]

    Lee, K.O., Mai, K.M., Park, S.: Green space accessibility helps buffer declined mental health during the covid-19 pandemic: evidence from big data in the united kingdom. Nat. Mental Health 1, 124–134 (2023)

    work page 2023

  6. [6]

    Environmental Health Perspectives 116, 196–202 (2008)

    Beelen, R., et al.: Long-term effects of traffic-related air pollution on mortality in a dutch cohort (nlcs-air study). Environmental Health Perspectives 116, 196–202 (2008)

    work page 2008

  7. [7]

    International Journal of Hygiene and Environmental Health 220, 1074–1080 (2017) 23

    Vert, C., et al.: Effect of long-term exposure to air pollution on anxiety and depression in adults: A cross-sectional study. International Journal of Hygiene and Environmental Health 220, 1074–1080 (2017) 23

    work page 2017

  8. [8]

    Environmental Research 153, 150–160 (2017)

    Morales-Su´ arez-Varela, M., Peraita-Costa, I., Llopis-Gonz´ alez, A.: Systematic review of the association between particulate matter exposure and autism spectrum disorders. Environmental Research 153, 150–160 (2017)

    work page 2017

Show all 79 references
  1. [9]

    Psychiatry Research 272, 8–17 (2019)

    Roberts, S., et al.: Exploration of no2 and pm2.5 air pollution and mental health problems using high-resolution data in london-based children from a uk longitudinal cohort study. Psychiatry Research 272, 8–17 (2019)

    2019

  2. [10]

    The Lancet Planetary Health 4, 64–73 (2020)

    Antonsen, S., et al.: Exposure to air pollution during childhood and risk of devel- oping schizophrenia: a national cohort study. The Lancet Planetary Health 4, 64–73 (2020)

    2020

  3. [11]

    Ji, J.S.: Air pollution and cardiovascular disease onset: hours, days, or years? The Lancet Public Health 7, 890–891 (2022)

    2022

  4. [12]

    The Lancet Regional Health - Western Pacific 24, 100463 (2022)

    Hu, Y., Ji, J.S., Zhao, B.: Restrictions on indoor and outdoor no2 emissions to reduce disease burden for pediatric asthma in china: A modeling study. The Lancet Regional Health - Western Pacific 24, 100463 (2022)

    2022

  5. [13]

    Nature 601, 380–387 (2022)

    Cooper, M.J., et al.: Global fine-scale changes in ambient no2 during covid-19 lockdowns. Nature 601, 380–387 (2022)

    2022

  6. [14]

    Nature Geosci 7, 792–795 (2014)

    Reuter, M., et al.: Decreasing emissions of nox relative to co2 in east asia inferred from satellite observations. Nature Geosci 7, 792–795 (2014)

    2014

  7. [15]

    Sci Rep 6, 35912 (2016)

    Foy, B., Lu, Z., Streets, D.G.: Satellite no2 retrievals suggest china has exceeded its nox reduction goals from the twelfth five-year plan. Sci Rep 6, 35912 (2016)

    2016

  8. [16]

    Sci Rep 4, 5887 (2015)

    Cuevas, C.A., et al.: Evolution of no2 levels in spain from 1996 to 2012. Sci Rep 4, 5887 (2015)

    2015

  9. [17]

    Nature 585, 225–233 (2020)

    Wells, K.C., et al.: Satellite isoprene retrievals constrain emissions and atmo- spheric oxidation. Nature 585, 225–233 (2020)

    2020

  10. [18]

    Sci Rep 10, 10066 (2020)

    Stavrakou, T., M¨ uller, J.-F., Bauwens, M., Boersma, K.F., Geffen, J.: Satellite evidence for changes in the no2 weekly cycle over large cities. Sci Rep 10, 10066 (2020)

    2020

  11. [19]

    Sci Rep 11, 1805 (2021)

    Shams, S.R., Jahani, A., Kalantary, S., Moeinaddini, M., Khorasani, N.: Artificial intelligence accuracy assessment in no2 concentration forecasting of metropolises air. Sci Rep 11, 1805 (2021)

    2021

  12. [20]

    Science 366, 723–727 (2019)

    Laughner, J.L., Cohen, R.C.: Direct observation of changing nox lifetime in north american cities. Science 366, 723–727 (2019)

    2019

  13. [21]

    Beirle, S., et al.: Pinpointing nitrogen oxide emissions from space. Sci. Adv. 5, 9800 (2019) 24

    2019

  14. [22]

    npj Urban Sustain 1, 35 (2021)

    Badia, A., et al.: A take-home message from covid-19 on urban air pollution reduction through mobility limitations and teleworking. npj Urban Sustain 1, 35 (2021)

    2021

  15. [23]

    Nat Commun 12, 243 (2021)

    Song, W., et al.: Important contributions of non-fossil fuel nitrogen oxides emissions. Nat Commun 12, 243 (2021)

    2021

  16. [24]

    Nature Geosci 10, 914–918 (2017)

    Grange, S.K., Lewis, A.C., Moller, S.J., Carslaw, D.C.: Lower vehicular primary emissions of no2 in europe than assumed in policy projections. Nature Geosci 10, 914–918 (2017)

    2017

  17. [25]

    Cities 96, 102481 (2020)

    Ibrahim, M.R., Haworth, J., Cheng, T.: Understanding cities with machine eyes: A review of deep computer vision in urban analytics. Cities 96, 102481 (2020)

    2020

  18. [26]

    IEEE Transactions on Intelligent Transportation Systems 25(3), 2827–2836 (2024) https://doi.org/10.1109/TITS

    Ma, C., Song, J., Xu, Y., Fan, H., Wu, X., Sun, T.: Vehicle-based machine vision approaches in intelligent connected system. IEEE Transactions on Intelligent Transportation Systems 25(3), 2827–2836 (2024) https://doi.org/10.1109/TITS. 2023.3276325

    2024

  19. [27]

    Neural Comput & Applic 35, 25097–25106 (2023) https://doi.org/10.1007/s00521-023-08712-9

    Ma, C., Song, J., Xu, Y.e.a.: Reducing environment exposure to covid-19 by iot sensing and computing with deep learning. Neural Comput & Applic 35, 25097–25106 (2023) https://doi.org/10.1007/s00521-023-08712-9

    2023 doi

  20. [28]

    ACS ES&T Engineering 3(1) (2022) https://doi.org/10.1021/acsestengg.2c00248

    Song, J., Fan, H., Gao, M., Xu, Y., Ran, M., Liu, X., Guo, Y.: Toward high- performance map-recovery of air pollution using machine learning. ACS ES&T Engineering 3(1) (2022) https://doi.org/10.1021/acsestengg.2c00248

    2022 doi

  21. [29]

    International Journal of Geographical Information Science 30(5), 881–906 (2016)

    Si la-Nowicka, K., Vandrol, J., Oshan, T., Long, J.A., Demˇ sar, U., Fotheringham, A.S.: Analysis of human mobility patterns from gps trajectories and contextual information. International Journal of Geographical Information Science 30(5), 881–906 (2016)

    2016

  22. [30]

    Nature 587(7834), 402–407 (2020)

    Alessandretti, L., Aslak, U., Lehmann, S.: The scales of human mobility. Nature 587(7834), 402–407 (2020)

    2020

  23. [31]

    Nature Human Behaviour 4(8), 800–810 (2020)

    Kraemer, M.U., Sadilek, A., Zhang, Q., Marchal, N.A., Tuli, G., Cohn, E.L., Hswen, Y., Perkins, T.A., Smith, D.L., Reiner Jr, R.C., et al.: Mapping global variation in human mobility. Nature Human Behaviour 4(8), 800–810 (2020)

    2020

  24. [32]

    Environmental pollution 229, 496–504 (2017)

    Gately, C.K., Hutyra, L.R., Peterson, S., Wing, I.S.: Urban emissions hotspots: Quantifying vehicle congestion and air pollution using mobile phone gps data. Environmental pollution 229, 496–504 (2017)

    2017

  25. [33]

    TfL London cameras (2021)

    2021

  26. [34]

    Lecture 25 Notes in Mathematics, vol

    Lyons, T.J., Caruana, M., L´ evy, T.: Differential Equations Driven by Rough Paths: ´Ecole d’ ´Et´ e de Probabilit´ es de Saint-Flour XXXIV-2004, 1st edn. Lecture 25 Notes in Mathematics, vol. 1908. Springer, ??? (2007). https://doi.org/10.1007/ 978-3-540-45886-2 . Part of the...

    2007 doi

  27. [35]

    Accessed: 2023-02-12 (2014)

    Lyons, T.: Rough paths, Signatures and the modelling of functions on streams. Accessed: 2023-02-12 (2014)

    2014

  28. [36]

    Oxford University Press, ??? (2002)

    Lyons, T., Qian, Z.: System Control and Rough Paths. Oxford University Press, ??? (2002). https://doi.org/10.1093/acprof:oso/9780198506485.001.0001 . Accessed: 2023-02-12

    2002

  29. [37]

    Atmospheric Chemistry and Physics 20(3), 1483–1495 (2020)

    Shah, V., Jacob, D.J., Li, K., Silvern, R.F., Zhai, S., Liu, M., Lin, J., Zhang, Q.: Effect of changing no x lifetime on the seasonality and long-term trends of satellite-observed tropospheric no 2 columns over china. Atmospheric Chemistry and Physics 20(3), 1483–1495 (2020)

    2020

  30. [38]

    Analytical Chemistry 73(22), 5485–5493 (2001)

    Matsumi, Y., Murakami, S.-i., Kono, M., Takahashi, K., Koike, M., Kondo, Y.: High-sensitivity instrument for measuring atmospheric no2. Analytical Chemistry 73(22), 5485–5493 (2001)

    2001

  31. [39]

    Annual review of earth and planetary sciences 7(1), 443–472 (1979)

    Crutzen, P.J.: The role of no and no2 in the chemistry of the troposphere and stratosphere. Annual review of earth and planetary sciences 7(1), 443–472 (1979)

    1979

  32. [40]

    Geophysical Research Letters 31(23) (2004)

    Richter, A., Eyring, V., Burrows, J.P., Bovensmann, H., Lauer, A., Sierk, B., Crutzen, P.J.: Satellite measurements of no2 from international shipping emissions. Geophysical Research Letters 31(23) (2004)

    2004

  33. [41]

    Atmospheric Environment 30(23), 4007–4020 (1996)

    Mentel, T.F., Bleilebens, D., Wahner, A.: A study of nighttime nitrogen oxide oxidation in a large reaction chamber—the fate of no2, n2o5, hno3, and o3 at different humidities. Atmospheric Environment 30(23), 4007–4020 (1996)

    1996

  34. [42]

    Atmospheric Chemistry and Physics 16(8), 5283–5298 (2016)

    Liu, F., Beirle, S., Zhang, Q., D¨ orner, S., He, K., Wagner, T.: No x lifetimes and emissions of cities and power plants in polluted background estimated by satellite observations. Atmospheric Chemistry and Physics 16(8), 5283–5298 (2016)

    2016

  35. [43]

    Health Effects Institute (2011)

    Kelly, F., et al.: The London Low Emission Zone Baseline Study. Health Effects Institute (2011)

    2011

  36. [44]

    The MIT Press, ??? (2017)

    Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. The MIT Press, ??? (2017)

    2017

  37. [45]

    Comput Soc Netw 6, 11 (2019)

    Zhang, S., Tong, H., Xu, J., Maciejewski, R.: Graph convolutional networks: a comprehensive review. Comput Soc Netw 6, 11 (2019)

    2019

  38. [46]

    Accessed: 2023-02-12 (2018)

    Veliˇ ckovi´ c, P., et al.: Graph Attention Networks. Accessed: 2023-02-12 (2018)

    2018

  39. [47]

    Hamilton, W., Ying, Z., Leskovec, J.: Inductive Representation Learning on Large Graphs (Published date unknown) 26

  40. [48]

    In: Proceedings of the 2005 IEEE International Joint Conference on Neural Networks, pp

    Gori, M., Monfardini, G., Scarselli, F.: A new model for learning in graph domains. In: Proceedings of the 2005 IEEE International Joint Conference on Neural Networks, pp. 729–734 (2005). IEEE

    2005

  41. [49]

    IEEE Signal Process

    Bronstein, M.M., Bruna, J., LeCun, Y., Szlam, A., Vandergheynst, P.: Geometric deep learning: going beyond euclidean data. IEEE Signal Process. Mag. 34, 18–42 (2017)

    2017

  42. [50]

    Annals of the American Association of Geographers 112, 2152–2173 (2022)

    Robinson, C., Franklin, R.S., Roberts, J.: Optimizing for equity: Sensor cover- age, networks, and the responsive city. Annals of the American Association of Geographers 112, 2152–2173 (2022)

    2022

  43. [51]

    In: 2017 IEEE Con- ference on Computer Vision and Pattern Recognition (CVPR), pp

    Redmon, J., Farhadi, A.: Yolo9000: Better, faster, stronger. In: 2017 IEEE Con- ference on Computer Vision and Pattern Recognition (CVPR), pp. 6517–6525 (2017). IEEE

    2017

  44. [52]

    Redmon, J., Farhadi, A.: YOLOv3: An Incremental Improvement (2018)

    2018

  45. [53]

    Ultralytics: YOLOv5 (2021)

    2021

  46. [54]

    nwojke: DeepSort (2019)

    2019

  47. [55]

    In: 2016 IEEE International Conference on Image Processing (ICIP), pp

    Bewley, A., Ge, Z., Ott, L., Ramos, F., Upcroft, B.: Simple online and realtime tracking. In: 2016 IEEE International Conference on Image Processing (ICIP), pp. 3464–3468 (2016). IEEE

    2016

  48. [56]

    In: Computer Vision – ECCV 2014, pp

    Lin, T.-Y., et al.: Microsoft coco: Common objects in context. In: Computer Vision – ECCV 2014, pp. 740–755. Springer, ??? (2014)

    2014

  49. [57]

    In: Proceedings of the Seventh IEEE International Conference on Computer Vision, vol

    Zhang, Z.: Flexible camera calibration by viewing a plane from unknown orienta- tions. In: Proceedings of the Seventh IEEE International Conference on Computer Vision, vol. 1, pp. 666–673 (1999). IEEE

    1999

  50. [58]

    arXiv:2412.16229 [cs.CV]

    Ibrahim, M.R.: TopView: Vectorising road users in a bird’s eye view from uncalibrated street-level imagery with deep learning. arXiv:2412.16229 [cs.CV]. Submitted on 18 Dec 2024. 28 pages (2024). https://doi.org/10.48550/arXiv.2412. 16229

    2024 doi arXiv

  51. [59]

    Published source unknown 3, 4 (2012)

    Venkatesh, M., Vijayakumar, P.: A simple bird’s eye view transformation technique. Published source unknown 3, 4 (2012)

    2012

  52. [60]

    In: 2019 IEEE 5th International Conference on Wireless and Telematics (ICWT), pp

    Mardiati, R., Mulyana, E., Maryono, I., Usman, K., Priatna, T.: The derivation of matrix transformation from pixel coordinates to real-world coordinates for vehicle trajectory tracking. In: 2019 IEEE 5th International Conference on Wireless and Telematics (ICWT), pp. 1–5 (2019). IEEE

    2019

  53. [61]

    Sensors 10, 2027–2044 (2010) 27

    Escalera, A., Armingol, J.M.: Automatic chessboard detection for intrinsic and extrinsic camera parameter calibration. Sensors 10, 2027–2044 (2010) 27

    2010

  54. [62]

    Journal of Financial Econometrics 8, 193–243 (2010)

    White, H., Lu, X.: Granger causality and dynamic structural systems. Journal of Financial Econometrics 8, 193–243 (2010)

    2010

  55. [63]

    In: Proceedings of the 2012 SIAM International Conference on Data Mining, pp

    Bahadori, M.T., Liu, Y.: Granger causality analysis in irregular time series. In: Proceedings of the 2012 SIAM International Conference on Data Mining, pp. 660–671 (2012). Society for Industrial and Applied Mathematics

    2012

  56. [64]

    (eds.) Causal Inference in Time Series Analysis, pp

    Eichler, M.: In: Berzuini, C., Dawid, P., Bernardinelli, L. (eds.) Causal Inference in Time Series Analysis, pp. 327–354. John Wiley & Sons, Ltd, ??? (2012)

    2012

  57. [65]

    Science 338, 496–500 (2012)

    Sugihara, G., et al.: Detecting causality in complex ecosystems. Science 338, 496–500 (2012)

    2012

  58. [66]

    Journal of the American Statistical Association 77, 304–313 (1982)

    Geweke, J.: Measurement of linear dependence and feedback between multiple time series. Journal of the American Statistical Association 77, 304–313 (1982)

    1982

  59. [67]

    Journal of Econometrics 21, 161–194 (1983)

    Geweke, J., Meese, R., Dent, W.: Comparing alternative tests of causality in temporal systems. Journal of Econometrics 21, 161–194 (1983)

    1983

  60. [68]

    IEEE Transactions on Systems, Man, and Cybernetics SMC-6, 325–327 (1976)

    Dudani, S.A.: The distance-weighted k-nearest-neighbor rule. IEEE Transactions on Systems, Man, and Cybernetics SMC-6, 325–327 (1976)

    1976

  61. [69]

    Geographical Analysis 27, 93–115 (1995)

    Anselin, L.: Local indicators of spatial association—lisa. Geographical Analysis 27, 93–115 (1995)

    1995

  62. [70]

    Geographical Analysis 24, 189–206 (2010)

    Getis, A., Ord, J.K.: The analysis of spatial association by use of distance statistics. Geographical Analysis 24, 189–206 (2010)

    2010

  63. [71]

    (ed.) Misspecification in Linear Spatial Regression Models, pp

    Florax, R.J.G.M., Nijkamp, P.: In: Kempf-Leonard, K. (ed.) Misspecification in Linear Spatial Regression Models, pp. 695–707. Elsevier, ??? (2005)

    2005

  64. [72]

    9393–9399

    Rey, S.J.: Mathematical Models in Geography, pp. 9393–9399. Pergamon, ??? (2001)

    2001

  65. [73]

    Accessed: 2023-02-12 (2016)

    Chevyrev, I., Kormilitzin, A.: A Primer on the Signature Method in Machine Learning. Accessed: 2023-02-12 (2016)

    2016

  66. [74]

    The London Lorry Control Scheme (LLCS) (1985)

    1985

  67. [75]

    In: Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference On, pp

    Dahl, G.E., Sainath, T.N., Hinton, G.E.: Improving deep neural networks for lvcsr using rectified linear units and dropout. In: Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference On, pp. 8609–8613 (2013). IEEE

    2013

  68. [76]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pp

    Ibrahim, M.R., Lyons, T.: Imagesig: A signature transform for ultra-lightweight image recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pp. 3649–3659 (2022) 28

    2022

  69. [77]

    The Journal of Machine Learning Research 15, 1929–1958 (2014)

    Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: A simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research 15, 1929–1958 (2014)

    2014

  70. [78]

    Glorot, X., Bengio, Y.: Understanding the difficulty of training deep feedforward neural networks

  71. [79]

    https:// arxiv.org/abs/1412.6980

    Kingma, D.P., Ba, J.: Adam: A Method for Stochastic Optimization. https:// arxiv.org/abs/1412.6980. Accessed: 2019-04-23 (2015) Data availability All raw data sources are listed in the Materials and Methods section. Supplementary Materials S1 - S2 Figures S1 - S5Tables 29 Fig....

    2015 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.