Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Towards Scalable Insect Monitoring: Ultra-Lightweight CNNs as On-Device Triggers for Insect Camera Traps

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read Ultra-lightweight CNNs can replace PIR sensors as triggers for insect camera traps.

desk verdict A solid, honestly-reported TinyML trigger for close-range insect traps; the Pollinator-Detect failure is in the abstract, but the conclusion still overreaches toward a generic trap. read the letter →

arxiv 2411.14467 v2 pith:34GE6ZLA submitted 2024-11-18 q-bio.QM cs.CVcs.LGeess.IV

classification q-bio.QMcs.CVcs.LGeess.IV
keywords ArtificialIntelligenceComputationalEntomologyConservationTechnologyCameraTrapsInsectDeclinesTinyMLBiodiversityInsects
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

Camera traps for insects have a trigger problem: passive infrared sensors cannot detect small, cold-blooded animals, so traps either miss insects or fill memory with blanks. This paper proposes replacing the sensor entirely with an ultra-lightweight binary CNN that runs continuously on the camera's microcontroller and decides, frame by frame, whether an insect is present. On test data the eight model variants reach 91.8–96.4% AUROC; on field data from distributions unseen in training, most stay above 87%, and the best model still pulls 87.2%. Deployed on an ESP32-S3, the trigger draws less than 300 mW, processes two to eight frames per second, and runs for weeks on a single battery. The unresolved edge is scale: when insects are small in a wide field of view, as in the Pollinator-Detect dataset, AUROC drops to 0.588–0.691, so the trigger works only where camera geometry keeps subjects large enough.

What carries the argument

The load-bearing mechanism is a MobileNetv2 binary classifier, shrunk with width multiplier α ∈ {0.1, 0.35} and small input tensors (96×96 or 120×160), then post-training quantised to 8-bit integers for microcontroller execution. Tiling splits high-resolution field frames into quarters or eighths so insects retain enough pixels at the model's low input resolution, and a single thresholded output neuron decides whether to save the full-resolution frame. Specificity of this threshold governs storage efficiency; recall governs responsiveness.

What would settle it

Deploy model 8 on an ESP32-S3 with a wide-angle lens so that insects typically occupy under 5% of the frame, run it on manually labelled field footage, and compare its AUROC with the >0.87 out-of-domain figure; if it falls toward 0.588, as on Pollinator-Detect, the trigger is only viable for close-range or bait-attracted camera setups.

Watch

Extended reading notes

Core claim

Ultra-lightweight binary classifiers can act as the trigger in an insect camera trap, replacing PIR sensors entirely. The paper trains eight MobileNetv2 variants—94,449 to 411,489 parameters, inputs from 96×96 greyscale to 120×160 RGB—on a balanced mix of iNaturalist, Ecostack, and Insect-Detect images, and shows they distinguish insects from background with 91.8–96.4% test AUROC and strong out-of-domain transfer on AMI-CT. A tiling strategy preserves small insect detail when field images are downscaled, and post-training int8 quantisation lets the models run on an ESP32-S3 at 0.3–8 frames per second under 300 mW. The decisive finding is the failure mode: on Pollinator-Detect, where insects occupy few pixels in wide field-of-view images, AUROC collapses to as low as 0.588, so the method's reach is bounded by the deployed camera's geometry.

Load-bearing premise

The trigger only works if insects remain discriminable at the model's small input resolution after downscaling or tiling; when an insect occupies only a few pixels in a wide field of view, the classifier cannot separate it from background, as the Pollinator-Detect results show.

Editorial extensions

If this is right

  • Existing insect camera traps can replace PIR or continuous recording with this trigger and extend battery life to 15–37 days on a 110-Wh battery.
  • Zero trigger-to-capture latency means fast-moving insects are recorded at the moment of detection, not after a sensor delay.
  • High specificity saves only frames containing insects, cutting the blank-image waste that plagues PIR-triggered traps.
  • The released models and codebase let other groups retrain the trigger as new field datasets accumulate.
  • The same architecture could trigger traps for other small ectotherms that infrared sensors miss.

Reading between the lines

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

  • Editorial inference: the Pollinator-Detect results imply the real deployment envelope is close-range or lure-based camera geometry where the insect spans a large part of the frame; system designers should assume the trigger degrades, not fails gracefully, as subject size shrinks.
  • Editorial inference: because a tile is labelled positive only if it contains more than half of a bounding box, an insect straddling a tile boundary can be missed entirely; an overlapping-tile scheme is a natural testable extension.
  • Editorial inference: at 2–8 frames per second, very fast fliers may cross the field of view between frames; coupling the CNN with a cheap motion wake-up could recover those detections without losing the zero-latency advantage.
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

4 major / 6 minor

Summary. This manuscript proposes replacing PIR triggers in insect camera traps with ultra-lightweight binary MobileNetV2 classifiers running on an ESP32-S3 microcontroller. The authors train eight model variants (two width multipliers, two input resolutions, grayscale/RGB) on a combined dataset of iNaturalist 2017 insect and plant images plus tiled Ecostack and Insect-Detect field data, then evaluate on a held-out test set, two in-distribution field datasets, and two unseen field datasets (Pollinator-Detect and AMI-CT). They report test AUROC values of 0.918-0.964, strong results on AMI-CT (0.818-0.872), but markedly weaker out-of-domain results on Pollinator-Detect (0.588-0.691). The paper also presents saliency-map analyses and on-device power/latency measurements, claiming a maximum power draw below 300 mW and zero trigger-to-capture latency. The authors provide model weights, code, and documentation for community use.

Significance. If the claims hold, this is a useful step toward low-cost, long-deployment insect monitoring, particularly for close-range or lure-based setups where the insect occupies a substantial part of the frame. The strengths of the paper are its large-scale training data, systematic exploration of model variants, transparent reporting of weak out-of-domain results in the v2 abstract, and the inclusion of quantisation and on-device measurements. The contribution is not a generic insect camera-trap trigger, however: the Pollinator-Detect results show that the models can fail badly in unconstrained field views, and the power/FPS comparison is not yet system-level. These issues are fixable by re-scoping the claims and tightening the evaluation protocol.

major comments (4)
  1. [§3, Table 3; Abstract; §4.1] The out-of-domain Pollinator-Detect AUROC values (0.588-0.691 across models 1-8) are close to chance, and the v2 abstract correctly reports these as the lower end of the field-data range. However, Section 4.1 states that 'the models are excellent candidates for insect triggers' without the deployment-geometry caveat that Section 4.3 acknowledges. The paper should either explicitly restrict the headline claim to close-range, lure-based, or tiled high-resolution deployments, or provide a quantitative deployment envelope (e.g., minimum insect apparent size or maximum camera distance for which the models retain high AUROC). As written, the generic 'insect camera trap trigger' claim is not supported by the Pollinator-Detect results.
  2. [§2.7, Figure 5, Table 4] The power and FPS measurements are taken on a minimal firmware that captures images at the model input size and runs inference; the tiling pipeline of Section 2.2, full-resolution capture, SD-card writes, and any sleep-state scheduling are not included. Table 4 nevertheless presents this as a system-level comparison with complete monitoring devices. Please report an end-to-end loop measurement (including tiling if it is part of the deployment claim) or re-label the 'Ours' row as inference-only and adjust the battery-life calculation to reflect the missing system components.
  3. [§2.2, Table 2, Table 3] The tiling protocol discards all tiles whose bounding boxes overlap the tile by less than 50% of the bounding-box area. This removes exactly the ambiguous, partially visible insects that are common in real deployments, so the in-distribution accuracy and AUROC in Table 3 are likely optimistic upper bounds. Please quantify the number of discarded tiles per dataset and report sensitivity to the overlap threshold (e.g., tile-positive-if-any-overlap or soft labels) so readers can judge how much the 50% cutoff affects the headline numbers.
  4. [§2.6, Equations 1-3, Figure 4] The saliency metric \bar{P}^t is defined as the average proportion of salient pixels inside insect bounding boxes, but the paper does not state how negative (background) images are handled in this average; if negatives contain no bounding boxes, the per-image quantity is undefined. Additionally, the metric rewards models that concentrate all saliency on a single insect pixel, so it should be supplemented with a coverage measure (the fraction of insect pixels that are salient) before claiming that the learned representation is robust. Please clarify the denominator and report both precision and coverage of saliency.
minor comments (6)
  1. [Abstract] The phrase 'minimal saving of false positive images' should be 'minimal storage of false positive images' for clarity.
  2. [Table 1] The parameter count range in the caption says '94,449-441,489', but the table values max at 411,489; this appears to be a typo.
  3. [§2.1] The data are promised for Dryad upon acceptance, but no review-stage access is provided. Since the code is attached, please also provide a review link to the datasets to enable reproduction.
  4. [Figure 5] No error bars or repeated-measurement statistics are shown for the power/latency data; please state the number of trials and the precision of the ampere meter.
  5. [§2.7] The company name is misspelled as 'Espressive'; it should be 'Espressif'.
  6. [§4.1] The text refers to 'AMT-CT'; elsewhere the dataset is called 'AMI-CT'. Please use consistent notation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; accuracy, latency, and power claims rest on empirical evaluation and direct measurement, not on self-referential derivation.

full rationale

The paper's central claims are supported by standard empirical procedures rather than by circular construction. Model parameters are learned from training data (iNat2017-insecta, Ecostack, Insect-Detect) and evaluated on held-out test splits and external field datasets (Pollinator-Detect, AMI-CT) using AUROC, accuracy, precision, F1, and specificity (Table 3). No fitted parameter is renamed as a prediction: the reported AUROC values are direct evaluations on data not used to fit the corresponding thresholds or weights. The 'zero latency' claim is an architectural property of the proposed capture-then-classify pipeline, not a numerical result derived from the definition of latency; the paper explicitly explains that images are captured continuously and decisions are made retrospectively, so there is no trigger-to-capture delay by design. Power consumption and frame-rate results are measured on an ESP32-S3 test firmware with a bench supply (Section 2.7, Figure 5), not derived from the model definitions. The only self-citation (Riechmann et al., reference [46]) appears in contextual discussion about motion vectors and distributing inferences across frames; it is not load-bearing for the paper's central claims. The low Pollinator-Detect AUROC values (0.588-0.691) weaken the generality of the 'unseen distribution' statement, but this is a scope or correctness concern, not a circularity. No equation in the paper reduces to its own input, and no claim is justified solely by an unverified self-citation.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard supervised learning assumptions plus a few paper-specific choices: the heuristic model width, input size, and tiling threshold. No new entities or physical mechanisms are introduced. The saliency and power measurements carry unexamined assumptions about what the metrics mean in the field.

free parameters (5)
  • alpha (width multiplier) values = 0.1, 0.35
    Chosen heuristically to trade off model size and accuracy; directly sets the parameter count and the reported results.
  • Input spatial size and color channels = 96x96x1, 96x96x3, 120x160x1, 120x160x3
    Input configurations were selected by heuristic search; smaller inputs reduce compute but also reduce detection detail.
  • Tiling label overlap threshold = 50%
    Tiles with bounding boxes covering less than 50% are discarded; this choice alters the training set and the reported AUROC.
  • Dropout rate and learning rate = 0.5, 0.001
    Standard regularization and optimizer settings chosen without an ablation, so their influence on the results is untested.
  • Tiling parameter n for evaluation = 2 (quarter-size tiles)
    Field datasets are tiled with n=2 for the reported AUROC; other tile sizes would change results.
assumptions (5)
  • domain assumption Pre-training on iNaturalist 2021-mini confers transferable visual features.
    Invoked in Section 2.3; the claim that transfer learning improves accuracy assumes the source dataset helps the target task.
  • domain assumption Source dataset bounding boxes and labels are correct.
    Used in Section 2.2 to assign binary labels to tiles; labeling errors would train the model on mislabeled data.
  • domain assumption Saliency gradient overlap with bounding boxes measures model trustworthiness.
    Section 2.6; vanilla gradients are known to be noisy, and no random baseline is provided, so the metric's validity is assumed.
  • domain assumption A 50% bounding box overlap threshold for positive tiles does not bias the model.
    Section 2.2; discarding tiles with partial insects removes hard examples and may inflate measured performance.
  • domain assumption The desk ammeter readings at 3.3 V reflect real deployment power draw.
    Section 2.7; the power consumption claim rests on this single measurement setup.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Scalable Insect Monitoring: Ultra-Lightweight CNNs as On-Device Triggers for Insect Camera Traps." pith.science (2026). https://pith.science/paper/34GE6ZLA

@misc{pith2026241114467,
  author       = {Pith},
  title        = {Pith review of: Towards Scalable Insect Monitoring: Ultra-Lightweight CNNs as On-Device Triggers for Insect Camera Traps},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/34GE6ZLA}},
  note         = {Machine review of arXiv:2411.14467}
}
read the original abstract

Camera traps, combined with AI, have emerged as a way to achieve automated, scalable biodiversity monitoring. However, the passive infrared (PIR) sensors that trigger camera traps are poorly suited for detecting small, fast-moving ectotherms such as insects. Insects comprise over half of all animal species and are key components of ecosystems and agriculture. The need for an appropriate and scalable insect camera trap is critical in the wake of concerning reports of declines in insect populations. This study proposes an alternative to the PIR trigger: ultra-lightweight convolutional neural networks running on low-powered hardware to detect insects in a continuous stream of captured images. We train a suite of models to distinguish insect images from backgrounds. Our design achieves zero latency between trigger and image capture. Our models are rigorously tested and achieve high accuracy ranging from 91.8% to 96.4% AUC on validation data and >87% AUC on data from distributions unseen during training. The high specificity of our models ensures minimal saving of false positive images, maximising deployment storage efficiency. High recall scores indicate a minimal false negative rate, maximising insect detection. Further analysis with saliency maps shows the learned representation of our models to be robust, with low reliance on spurious background features. Our system is also shown to operate deployed on off-the-shelf, low-powered microcontroller units, consuming a maximum power draw of less than 300mW. This enables longer deployment times using cheap and readily available battery components. Overall we offer a step change in the cost, efficiency and scope of insect monitoring. Solving the challenging trigger problem, we demonstrate a system which can be deployed for far longer than existing designs and budgets power and bandwidth effectively, moving towards a generic insect camera trap.

Figures

Figures reproduced from arXiv: 2411.14467 by the authors.

Figure 1
Figure 1. Software pipeline showcasing how the trigger (CNN, binary classifier) [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Simplified workflow for saliency map analysis showing the main stages. [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗
Figure 3
Figure 3. ROC curves for each model computed from the test dataset. Shows true [PITH_FULL_IMAGE:figures/full_fig_p022_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Resultant curves for our saliency score, [PITH_FULL_IMAGE:figures/full_fig_p024_4.png]
Figure 5
Figure 5. Figure 5: Average power consumption for each model versus operational frames per [PITH_FULL_IMAGE:figures/full_fig_p025_5.png]
Figure 6
Figure 6. Figure 6: Example predictions from our test dataset, executed using model 8. Input [PITH_FULL_IMAGE:figures/full_fig_p026_6.png]
Figure 7
Figure 7. Figure 7: Example predictions from datasets unseen in training, executed using [PITH_FULL_IMAGE:figures/full_fig_p027_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. EV-Flying: an Event-based Dataset for In-The-Wild Recognition of Flying Objects

    cs.CV 2025-06 conditional novelty 6.0 of 10

    EV-Flying is a hand-annotated event-camera dataset of birds, insects, and drones, with a PointNet++ benchmark reaching about 72% single-chunk and 92% full-track accuracy.

Reference graph

Works this paper leans on

58 extracted references · 32 canonical work pages · cited by 1 Pith paper

  1. [1]

    Nigel E. Stork. How many species of insects and other terrestrial arthropods are there on earth? Annual Review of Entomology , 63 (Volume 63, 2018):31–45, 2018. ISSN 1545-4487. doi: https://doi.org/10. 1146/annurev-ento-020117-043348. URL https://www.annualreviews.org/ content/journals/10.1146/annurev-ento-020117-043348

  2. [2]

    Theglobalemergence/resurgenceofarboviraldiseasesaspublic health problems

    DuaneJGubler. Theglobalemergence/resurgenceofarboviraldiseasesaspublic health problems. Archives of Medical Research, 33(4):330–342, 2002. ISSN 0188-4409. doi: https://doi.org/10.1016/S0188-4409(02)00378-8. URL https: //www.sciencedirect.com/science/article/pii/S0188440902003788

  3. [3]

    Bowler, Konstantin B

    Roel van Klink, Diana E. Bowler, Konstantin B. Gongalsky, Ann B. Swengel, Alessandro Gentile, and Jonathan M. Chase. Meta-analysis reveals declines in terrestrial but increases in freshwater insect abundances.Science, 368(6489): 33 417–420, 2020. doi: 10.1126/science.aax9931. URL https://www.science. org/doi/abs/10.1126/science.aax9931

  4. [4]

    Wagner, Eliza M

    David L. Wagner, Eliza M. Grames, Matthew L. Forister, May R. Berenbaum, and David Stopak. Insect decline in the anthropocene: Death by a thousand cuts. Proceedings of the National Academy of Sciences, 118(2):e2023989118,

  5. [5]

    Hallmann, Martin Sorg, Eelke Jongejans, Henk Siepel, Nick Hofland, Heinz Schwan, Werner Stenmans, Andreas Müller, Hubert Sumser, Thomas Hörren, Dave Goulson, and Hans de Kroon

    Caspar A. Hallmann, Martin Sorg, Eelke Jongejans, Henk Siepel, Nick Hofland, Heinz Schwan, Werner Stenmans, Andreas Müller, Hubert Sumser, Thomas Hörren, Dave Goulson, and Hans de Kroon. More than 75 percent decline over 27 years in total flying insect biomass in protected areas.PLOS ONE, 12(10): 1–21, 10 2017. doi: 10.1371/journal.pone.0185809. URLhttps:...

  6. [6]

    Montgomery, Michael W

    Graham A. Montgomery, Michael W. Belitz, Rob P. Guralnick, and Morgan W. Tingley. Standards and best practices for monitoring and benchmarking in- sects. Frontiers in Ecology and Evolution, 8, 2021. ISSN 2296-701X. doi: 10.3389/fevo.2020.579193. URL https://www.frontiersin.org/journals/ ecology-and-evolution/articles/10.3389/fevo.2020.579193

  7. [7]

    Rudolf Meier, Emily Hartop, Christian Pylatiuk, and Amrita Srivathsan. To- wards holistic insect monitoring: species discovery, description, identification and traits for all insects.Philosophical Transactions of the Royal Society B: Bi- 34 ological Sciences, 379(1904):20230120, 2024. doi: 10.1098/rstb.2023.0120. URL https://royalsocietypublishing.org/doi...

  8. [8]

    Han, Kush R

    Barbara A. Han, Kush R. Varshney, Shannon LaDeau, Ajit Subramaniam, Kathleen C. Weathers, and Jacob Zwart. A synergistic future for ai and ecol- ogy. Proceedings of the National Academy of Sciences, 120(38):e2220283120,

Show all 58 references
  1. [9]

    Høye, Johanna Ärje, Kim Bjerge, Oskar L

    Toke T. Høye, Johanna Ärje, Kim Bjerge, Oskar L. P. Hansen, Alexandros Iosifidis, Florian Leese, Hjalte M. R. Mann, Kristian Meissner, Claus Melvad, and Jenni Raitoharju. Deep learning and computer vision will transform ento- mology. Proceedings of the National Academy of Scie...

  2. [10]

    Scaling-up camera traps: monitoring the planet’s biodiversity with networks of remote sensors.Frontiers in Ecology and the Environment, 15(1):26–34, 2017

    Robin Steenweg, Mark Hebblewhite, Roland Kays, Jorge Ahumada, Jason T Fisher, Cole Burton, Susan E Townsend, Chris Carbone, J Marcus Rowcliffe, Jesse Whittington, Jedediah Brodie, J Andrew Royle, Adam Switalski, An- thony P Clevenger, Nicole Heim, and Lindsey N Rich. Scaling-u...

  3. [11]

    Palmer, Craig Packer, and Jeff Clune

    Mohammad Sadegh Norouzzadeh, Anh Nguyen, Margaret Kosmala, Alexandra 35 Swanson, Meredith S. Palmer, Craig Packer, and Jeff Clune. Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning.Proceedings of the National Academy of S...

  4. [12]

    URL https://www.pnas.org/doi/abs/ 10.1073/pnas.2002545117

    doi: 10.1073/pnas.2002545117. URL https://www.pnas.org/doi/abs/ 10.1073/pnas.2002545117

  5. [13]

    Welbourne, Andrew W

    Dustin J. Welbourne, Andrew W. Claridge, David J. Paull, and Andrew Lam- bert. How do passive infrared triggered camera traps operate and why does it matter? breaking down common misconceptions. Remote Sensing in Ecol- ogy and Conservation, 2(2):77–83, 2016. doi: https://doi.o...

  6. [14]

    Hobbs and Cheryl S

    Michael T. Hobbs and Cheryl S. Brehme. An improved camera trap for am- phibians, reptiles, small mammals, and large invertebrates.PLOS ONE, 12(10): 36 1–15, 10 2017. doi: 10.1371/journal.pone.0185026. URLhttps://doi.org/10. 1371/journal.pone.0185026

  7. [15]

    Whytock, Jędrzej Świeżewski, Joeri A

    Robin C. Whytock, Jędrzej Świeżewski, Joeri A. Zwerts, Tadeusz Bara-Słupski, Aurélie Flore Koumba Pambo, Marek Rogala, Laila Bahaa-el din, Kelly Boekee, Stephanie Brittain, Anabelle W. Cardoso, Philipp Henschel, David Lehmann, Brice Momboua, Cisquet Kiebou Opepa, Christopher O...

  8. [16]

    Wolff, Brenda Molano-Flores, and Jinelle H

    Qaim Naqvi, Patrick J. Wolff, Brenda Molano-Flores, and Jinelle H. Sperry. Camera traps are an effective tool for monitoring insect–plant interactions.Ecol- ogy and Evolution, 12(6):e8962, 2022. doi: https://doi.org/10.1002/ece3.8962. URL https://onlinelibrary.wiley.com/doi/ab...

  9. [17]

    Insect pest monitoring with camera-equipped traps: strengths and limitations.Journal of Pest Science, 94(2):203–217, 2021

    Michele Preti, François Verheggen, and Sergio Angeli. Insect pest monitoring with camera-equipped traps: strengths and limitations.Journal of Pest Science, 94(2):203–217, 2021. ISSN 1612-4766. doi: 10.1007/s10340-020-01309-4. URL https://doi.org/10.1007/s10340-020-01309-4

  10. [18]

    Ethicalissuesinlethalinsectmonitoring

    MeghanBarrettandBobFischer. Ethicalissuesinlethalinsectmonitoring. Cur- rent Opinion in Insect Science, 66:101279, 2024. ISSN 2214-5745. doi: https:// doi.org/10.1016/j.cois.2024.101279. URL https://www.sciencedirect.com/ science/article/pii/S2214574524001214

  11. [19]

    Houlihan, Mac Stone, Shawn E

    Peter R. Houlihan, Mac Stone, Shawn E. Clem, Mike Owen, and Thomas C. Emmel. Pollination ecology of the ghost orchid (Dendrophylax lindenii): A first description with new hypotheses for darwin’s orchids.Scientific Reports, 9(1):12850, 2019. ISSN 2045-2322. doi: 10.1038/s41598-...

  12. [20]

    URL https://zslpublications.onlinelibrary.wiley.com/doi/abs/ 10.1002/rse2.20

  13. [21]

    Genoud, Gregory M

    Topu Saha, Adrien P. Genoud, Gregory M. Williams, and Benjamin P. Thomas. Monitoring the abundance of flying insects and atmospheric conditions during a 9-month campaign using an entomological optical sensor.Scientific Reports, 13(1):15606, 2023. ISSN 2045-2322. doi: 10.1038/s...

  14. [22]

    Bioclip: A vision foundation model for the tree of life

    Samuel Stevens, Jiaman Wu, Matthew J Thompson, Elizabeth G Campolongo, Chan Hee Song, David Edward Carlyn, Li Dong, Wasila M Dahdul, Charles Stewart, Tanya Berger-Wolf, et al. Bioclip: A vision foundation model for the tree of life. InProceedings of the IEEE/CVF Conference on ...

  15. [23]

    Towards a dynamic vision sensor-based insect camera trap

    Eike Gebauer, Sebastian Thiele, Pierre Ouvrard, Adrien Sicard, and Benjamin Risse. Towards a dynamic vision sensor-based insect camera trap. InProceed- 37 ings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 7157–7166, 2024

  16. [24]

    The buz- zometer system: In situ audio recordings of pollinators in flight

    Paolo Volponi, Leonardo Dapporto, and Marta Skowron Volponi. The buz- zometer system: In situ audio recordings of pollinators in flight. Methods in Ecology and Evolution, 14(12):2985–2993, 2023. doi: https://doi.org/10. 1111/2041-210X.14224. URL https://besjournals.onlinelibra...

  17. [25]

    Christopher D

    J. Christopher D. Terry, Helen E. Roy, and Tom A. August. Thinking like a naturalist: Enhancing computer vision of citizen science images by harness- ing contextual data. Methods in Ecology and Evolution, 11(2):303–315, 2020. doi: https://doi.org/10.1111/2041-210X.13335. URL h...

  18. [26]

    Smith, and Vincent Deblauwe

    Vincent Droissart, Laura Azandi, Eric Rostand Onguene, Marie Savignac, Thomas B. Smith, and Vincent Deblauwe. Pict: A low-cost, modular, open- source camera trap system to study plant–insect interactions. Methods in Ecology and Evolution, 12(8):1389–1396, 2021. doi: https://do...

  19. [27]

    Insect-foundation: A foundation model and large-scale 1m dataset for visual insect understanding

    Hoang-Quan Nguyen, Thanh-Dat Truong, Xuan Bac Nguyen, Ashley Dowling, Xin Li, and Khoa Luu. Insect-foundation: A foundation model and large-scale 1m dataset for visual insect understanding. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pa...

  20. [28]

    Mann, Toke T

    Kim Bjerge, Quentin Geissmann, Jamie Alison, Hjalte M.R. Mann, Toke T. Høye, Mads Dyrmann, and Henrik Karstoft. Hierarchical classification of in- sects with multitask learning and anomaly detection. Ecological Informat- ics, 77:102278, 2023. ISSN 1574-9541. doi: https://doi.o...

  21. [29]

    An automated light trap to moni- tor moths (lepidoptera) using computer vision-based tracking and deep learn- ing

    Kim Bjerge, Jakob Bonde Nielsen, Martin Videbæk Sepstrup, Flemming Helsing-Nielsen, and Toke Thomas Høye. An automated light trap to moni- tor moths (lepidoptera) using computer vision-based tracking and deep learn- ing. Sensors, 21(2), 2021. ISSN 1424-8220. doi: 10.3390/s2102...

  22. [30]

    Kim Bjerge, Hjalte M. R. Mann, and Toke Thomas Høye. Real-time insect tracking and monitoring with computer vision and deep learning.Remote Sens- ing in Ecology and Conservation, 8(3):315–327, 2022. doi: https://doi.org/10. 1002/rse2.245. URL https://zslpublications.onlinelibr...

  23. [31]

    Insect detect: An open-source diy camera trap for automated insect monitoring

    Maximilian Sittinger, Johannes Uhler, Maximilian Pink, and Annette Herz. Insect detect: An open-source diy camera trap for automated insect monitoring. PLOS ONE, 19(4):1–28, 04 2024. doi: 10.1371/journal.pone.0295474. URL https://doi.org/10.1371/journal.pone.0295474

  24. [32]

    Kevin F. A. Darras, Marcel Balle, Wenxiu Xu, Yang Yan, Vincent G. Za- 39 kka, Manuel Toledo-Hernández, Dong Sheng, Wei Lin, Boyu Zhang, Zhen- zhong Lan, Li Fupeng, and Thomas C. Wanger. Eyes on nature: Embedded vision cameras for terrestrial biodiversity monitoring.Methods in ...

  25. [33]

    Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, and Hartwig Adam

    Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, and Hartwig Adam. Mobilenets: Ef- ficient convolutional neural networks for mobile vision applications. CoRR, abs/1704.04861, 2017. URL http://arxiv.org/abs/1704.04861

  26. [34]

    Visual wake words dataset, 2019

    Aakanksha Chowdhery, Pete Warden, Jonathon Shlens, Andrew Howard, and Rocky Rhodes. Visual wake words dataset, 2019. URLhttps://arxiv.org/ abs/1906.05721

  27. [35]

    Tinyml: Tools, applications, challenges, and future research directions

    Rakhee Kallimani, Krishna Pai, Prasoon Raghuwanshi, Sridhar Iyer, and Onel LA López. Tinyml: Tools, applications, challenges, and future research directions. Multimedia Tools and Applications, 83(10):29015–29045, 2024

  28. [36]

    dos Santos, Raimundo Barreto, and Juan G

    Fagner Cunha, Eulanda M. dos Santos, Raimundo Barreto, and Juan G. Colonna. Filtering empty camera trap images in embedded systems. InProceed- 40 ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 2438–2446, June 2021

  29. [37]

    Object detection of small insects in time-lapse camera recordings

    Kim Bjerge, Carsten Eie Frigaard, and Henrik Karstoft. Object detection of small insects in time-lapse camera recordings. Sensors, 23(16), 2023. ISSN 1424-8220. doi: 10.3390/s23167242. URLhttps://www.mdpi.com/1424-8220/ 23/16/7242. 41

  30. [38]

    McFarland, David Roy, Toke Thomas Høye, Maxim Larrivée, and David Rolnick

    Aditya Jain, Fagner Cunha, Michael James Bunsen, Juan Sebastián Cañas, Léonard Pasi, Nathan Pinoy, Flemming Helsing, JoAnne Russo, Marc Botham, Michael Sabourin, Jonathan Frechette, Alexandre Anctil, Yacksecari Lopez, Eduardo Navarro, Filonila Perez Pimentel, Ana Cecilia Zamor...

  31. [39]

    Mobilenetv2: Inverted residuals and linear bottlenecks

    Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, and Liang- Chieh Chen. Mobilenetv2: Inverted residuals and linear bottlenecks. InProceed- ings of the IEEE conference on computer vision and pattern recognition, pages 4510–4520, 2018

  32. [40]

    Kim Bjerge, Jamie Alison, Mads Dyrmann, Carsten Eie Frigaard, Hjalte M. R. Mann, and Toke Thomas Høye. Accurate detection and identification of in- sects from camera trap images with deep learning. PLOS Sustainability and Transformation, 2(3):1–18, 03 2023. doi: 10.1371/journa...

  33. [41]

    The inaturalist species classification and detection dataset

    Grant Van Horn, Oisin Mac Aodha, Yang Song, Yin Cui, Chen Sun, Alex Shep- ard, Hartwig Adam, Pietro Perona, and Serge Belongie. The inaturalist species classification and detection dataset. InProceedings of the IEEE conference on computer vision and pattern recognition, pages ...

  34. [42]

    Keras.https://keras.io, 2015

    François Chollet et al. Keras.https://keras.io, 2015

  35. [43]

    Howard, Hartwig Adam, and Dmitry Kalenichenko

    Benoit Jacob, Skirmantas Kligys, Bo Chen, Menglong Zhu, Matthew Tang, Andrew G. Howard, Hartwig Adam, and Dmitry Kalenichenko. Quantization and training of neural networks for efficient integer-arithmetic-only inference. CoRR, abs/1712.05877, 2017. URLhttp://arxiv.org/abs/1712.05877

  36. [44]

    Adam: A method for stochastic optimization

    Diederik Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In International Conference on Learning Representations (ICLR), San Diega, CA, USA, 2015

  37. [45]

    Tensorflow lite micro: Embedded ma- chine learning for tinyml systems

    Robert David, Jared Duke, Advait Jain, Vijay Janapa Reddi, Nat Jef- fries, Jian Li, Nick Kreeger, Ian Nappier, Meghna Natraj, Tiezhen Wang, Pete Warden, and Rocky Rhodes. Tensorflow lite micro: Embedded ma- chine learning for tinyml systems. In A. Smola, A. Dimakis, and I. Sto...

  38. [46]

    Motion vectors and deep neural networks for video camera traps

    Miklas Riechmann, Ross Gardiner, Kai Waddington, Ryan Rueger, Frederic Fol Leymarie, and Stefan Rueger. Motion vectors and deep neural networks for video camera traps. Ecological Informatics, 69:101657, 2022. ISSN 1574-

  39. [47]

    Benchmarking representation learning for natural world image collections, 2021

    Grant Van Horn, Elijah Cole, Sara Beery, Kimberly Wilber, Serge Belongie, 42 and Oisin Mac Aodha. Benchmarking representation learning for natural world image collections, 2021. URLhttps://arxiv.org/abs/2103.16483

  40. [48]

    Dropout: A simple way to prevent neural networks from over- fitting

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: A simple way to prevent neural networks from over- fitting. Journal of Machine Learning Research, 15(56):1929–1958, 2014. URL http://jmlr.org/papers/v15/srivastava14a.html

  41. [49]

    Interpretable Machine Learning

    Christoph Molnar. Interpretable Machine Learning. 2 edition, 2022. URL https://christophm.github.io/interpretable-ml-book

  42. [50]

    Enhancing lightweight neural networks for small object detection in iot applications

    Liam Boyle, Nicolas Baumann, Seonyeong Heo, and Michele Magno. Enhancing lightweight neural networks for small object detection in iot applications. In 2023 IEEE SENSORS, pages 01–04, 2023. doi: 10.1109/SENSORS56945.2023. 10325126

  43. [51]

    A survey on deep neural network pruning-taxonomy, comparison, analysis, and recommendations, 2024

    Hongrong Cheng, Miao Zhang, and Javen Qinfeng Shi. A survey on deep neural network pruning-taxonomy, comparison, analysis, and recommendations, 2024. URL https://arxiv.org/abs/2308.06767

  44. [52]

    Mandrekar

    Jayawant N. Mandrekar. Receiver operating characteristic curve in diagnostic test assessment. Journal of Thoracic Oncology, 5(9):1315–1316, 2010. ISSN 1556-0864. doi: https://doi.org/10.1097/JTO.0b013e3181ec173d. URL https: //www.sciencedirect.com/science/article/pii/S15560864...

  45. [53]

    Taylor, and Stefan C

    Stefan Schneider, Saul Greenberg, Graham W. Taylor, and Stefan C. Kremer. Three critical factors affecting automated image species recognition performance for camera traps. Ecology and Evolution, 10(7):3503–3517, 2020. doi: https: //doi.org/10.1002/ece3.6147. URL https://onlin...

  46. [57]

    User-based design specifications for the ultimate camera trap for wildlife research.Wildlife Research, 39(8):649–660, 2012

    PD Meek and A Pittet. User-based design specifications for the ultimate camera trap for wildlife research.Wildlife Research, 39(8):649–660, 2012

  47. [58]

    Snapshot serengeti, high-frequency annotated camera trap images of 40 mammalian species in an african savanna.Scientific Data, 2(1):150026, Jun 2015

    Alexandra Swanson, Margaret Kosmala, Chris Lintott, Robert Simpson, Arfon Smith, and Craig Packer. Snapshot serengeti, high-frequency annotated camera trap images of 40 mammalian species in an african savanna.Scientific Data, 2(1):150026, Jun 2015. ISSN 2052-4463. doi: 10.1038...

  48. [2021]

    URL https://www.pnas.org/doi/abs/ 10.1073/pnas.2023989118

    doi: 10.1073/pnas.2023989118. URL https://www.pnas.org/doi/abs/ 10.1073/pnas.2023989118

  49. [2023]

    URL https://www.pnas.org/doi/abs/ 10.1073/pnas.2220283120

    doi: 10.1073/pnas.2220283120. URL https://www.pnas.org/doi/abs/ 10.1073/pnas.2220283120

  50. [9541]

    URL https://www

    doi: https://doi.org/10.1016/j.ecoinf.2022.101657. URL https://www. sciencedirect.com/science/article/pii/S1574954122001066

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.