Pith. sign in

REVIEW 3 major objections 4 minor 46 references

MetaWild: A Multimodal Dataset for Animal Re-Identification with Environmental Metadata

T0 review · 3 major / 4 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 Environmental metadata encoded as text lifts animal re-identification by up to 6.5 mAP points on a new six-species benchmark.

desk verdict Valuable new dataset, but the central ReID evaluation is compromised by an unstated identity token in the prompt template. read the letter →

arxiv 2501.13368 v2 pith:KDNLGBVF submitted 2025-01-23 cs.CV cs.LG

Yuzhuo Li , Di Zhao , Tingrui Qiao , Yihao Wu , Bo Pang , Yun Sing Koh This is my paper · ORCID
classification cs.CVcs.LG
keywords Animalre-identificationVision-languagemodelEnvironmentalmetadataCameratrapDatasetfusionReIDMultimodallearning
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

Wildlife camera traps record far more than an animal's appearance: embedded overlays carry temperature, day/night status, and usable face-orientation cues. This paper argues that those environmental metadata are identity-discriminative and that existing animal re-identification (ReID) benchmarks throw them away. To test the idea, the authors build MetaWild, a 20,890-image, six-species dataset pairing each image with temperature, circadian rhythm, and face-orientation metadata, and they propose the Meta-Feature Adapter (MFA), a lightweight module that turns metadata into text prompts and fuses them into vision-language-model features. Across intra-species and cross-species protocols, adding MFA to CLIP-ReID and ReID-AW consistently improves mean average precision, with the largest intra-species gain being 6.5 points on Penguin. If the result holds, existing camera-trap deployments could get better individual identification without new hardware, only by using information already stamped on the images.

What carries the argument

The load-bearing object is the Meta-Feature Adapter (MFA), a lightweight plug-in that lets existing VLM-based ReID models consume metadata without architecture changes. MFA has two parts: feature experts (two MLP adapters with residual connections, one on the text branch and one on the image branch) and a gated cross-attention module. Metadata is rendered into the prompt template "A photo of a {species} {individual id} in {freezing...hot} temperature, with face direction {front/back/left/right}, captured during the {day/night}", embedded by the VLM text encoder, refined by the textual expert, and fused into image embeddings through cross-attention with a learned gating value $\gamma \in [0,1]$, giving $I_{\text{meta}} = \gamma A V + I'_x$. The gating is what lets the model suppress metadata that is irrelevant for a particular image; the residual connection guarantees the visual stream is always present. This mechanism is what carries the paper's reported gains.

What would settle it

Re-run the MFA evaluation with the {individual id} token replaced by a fixed generic word such as "animal" in both training and test prompts; if the mAP gains over visual-only baselines shrink to zero, the reported improvements came from leaking the identity label through the prompt. A second check is to rebuild query/gallery splits so no two frames of the same identity come from the same camera burst; if near-duplicate frames are currently shared across splits, background or timestamp cues could explain the match scores.

Watch

Extended reading notes

Core claim

The paper's central claim is that environmental metadata correlated with animal behavior—temperature, circadian rhythm, and face orientation—carry identity-discriminative signal that vision-only ReID models miss. The authors construct MetaWild from the NZ-TrailCams archive by cropping detected animals, assigning individual identities from temporal proximity plus visual verification, and extracting metadata from embedded overlays, then show that converting that metadata into fixed text prompts ("A photo of a Stoat in cool temperature, with face direction left, captured during the day") and injecting the resulting text embeddings into a VLM's visual feature stream yields consistent mAP improvements over visual-only baselines. Concretely, the Meta-Feature Adapter raises ReID-AW mAP from 48.8% to 55.3% on Penguin in intra-species evaluation, from 67.5% to 72.4% on Deer, and from 58.4% to 63.5% on Wallaby, with similar gains under leave-one-species-out cross-species evaluation. The intended upshot is that metadata integration is a cheap, effective axis of improvement for animal re-identification and that MetaWild provides the benchmark to measure it.

Load-bearing premise

The claim collapses if the {individual id} token in the metadata prompt is not masked at inference, because then the text encoder is handed the ground-truth label and the mAP gains measure label leakage rather than metadata value; the paper never states that this token is removed or replaced.

Editorial extensions

If this is right

  • Existing VLM-based ReID models can gain consistent, several-point mAP boosts by adding metadata, with no change to the visual backbone.
  • Camera-trap archives that carry overlay metadata can be repurposed for multimodal ReID without new field equipment.
  • Cross-species leave-one-domain-out results suggest metadata helps transfer identity features to unseen species, which matters when labels for a new species are scarce.
  • MetaWild gives the community a six-species benchmark for measuring how much of ReID accuracy can come from context rather than appearance.

Reading between the lines

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

  • A fair test of the paper's claim requires masking the identity token in the prompt; until then, the cleanest reading is that metadata helps, but part of the gain may come from the model reading the ground-truth label from the prompt.
  • If the identity-label leak is controlled and gains persist, metadata priors could be exploited even without appearance: individuals often have site, time, and temperature preferences, so a Bayesian ReID model that combines visual similarity with environmental priors might outperform the adapter.
  • The same prompt-and-adapter recipe could transfer to other wildlife archives with overlay metadata, making it a practical low-cost upgrade for conservation monitoring.
  • Because face orientation and day/night are metadata partially inferred from the image itself, the dataset cannot cleanly separate 'environmental' signal from appearance; a stronger test would use camera-logged metadata only, excluding manually annotated face direction.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper introduces MetaWild, a new animal re-identification (ReID) dataset constructed from NZ-TrailCams that pairs images with environmental metadata (temperature, circadian rhythm, face orientation), and proposes the Meta-Feature Adapter (MFA), a lightweight module to inject such metadata into VLM-based ReID models. The authors report intra-species and leave-one-domain-out inter-species experiments on six species, claiming that adding MFA consistently improves mAP and CMC-1 over visual-only baselines. The dataset and code are publicly released, and the experiments include 95% confidence intervals.

Significance. If the experimental protocol were sound, this work would address a real gap: existing animal ReID datasets are visual-only, and the idea of converting camera-trap metadata into text prompts for VLMs is useful and timely. The released dataset and the adapter design could be a valuable resource for the community. However, the central claim as written is not supported because the text prompt used in the MFA pipeline includes the ground-truth individual identity token, and the paper never states that this token is removed or masked at test time. Under a standard ReID protocol, this makes the reported gains potentially circular, so the current evidence for the value of environmental metadata is invalid.

major comments (3)
  1. [Section 4.1, Figure 4] The MFA prompt template is defined as "A photo of a {species} {individual id} in ..." and Figure 4 fills this with a concrete identity token ("<Bob>"). The paper never states that this token is dropped, replaced with a generic word, or randomized when prompts are built for query and gallery images at test time. Because the identity classification loss and the contrastive loss in Eq. (4) train the text branch to map identity tokens to discriminative embeddings, a test-time prompt containing the true identity gives the model access to the ground-truth label through the text modality; matching identical ID strings alone can solve the ReID task. The reported improvements (e.g., Table 2, Penguin ReID-AW+MFA mAP 48.8 to 55.3) are therefore uninterpretable as evidence that metadata helps. The authors must either explicitly specify a test-time protocol that removes the identity token, or rerun all experiments without identity information in the prompt.
  2. [Section 3.2] The identity annotation procedure assigns the same identity to animals captured "within narrow time windows (e.g., a few seconds) at the same camera location" and then refines groupings by visual inspection. The paper does not describe how the train/gallery/query splits are constructed with respect to time or camera bursts. If frames from the same burst appear in both query and gallery, matching can exploit near-identical backgrounds, lighting, and timestamps rather than individual identity, inflating all reported results. The authors should provide a temporal split (e.g., no query/gallery pairs within some minimum time gap) or at least report the distribution of time gaps between matched query and gallery images.
  3. [Section 5.1, Tables 2 and 3] The claim of "consistent improvements" is not accompanied by any ablation that isolates the effect of the identity token from the effect of the remaining metadata. Since the prompt contains identity, the gains in Tables 2 and 3 are exactly what identity leakage would produce, and the paper provides no experiment showing that temperature, day/night, or face orientation alone—without the identity token—yields the same improvement. A controlled comparison with the identity token removed from both training and test prompts is necessary to support the central claim.
minor comments (4)
  1. [Section 4.1] The temperature categories [freezing, cold, chilly, cool, warm, hot] are introduced without reporting the threshold values used to map numeric temperatures to these bins; this is needed for reproducibility.
  2. [Figure 4] The prompt example uses "<Bob>" as the identity token, while Section 3.2 describes numeric identities (e.g., "11"); the notation is inconsistent and should be unified.
  3. [Table 1 and throughout] The spelling of the species name is inconsistent: "Pūkeko" appears in some places and "P¯ukeko" in others; this should be standardized.
  4. [Section 5.1] CLIP-ZS is a zero-shot baseline while all MFA variants are fine-tuned on the training split, so the comparison is not controlled; the zero-shot row should be labeled as such and not treated as a same-protocol baseline.

Circularity Check

1 steps flagged · score 8.0 of 10

The MFA prompt template embeds the ground-truth {individual id} token; as written, the reported ReID gains can be obtained by reading the answer from the text input rather than from environmental metadata.

  1. self definitional [Section 4.1 'Feature Experts', prompt template and Figure 4]
    ""A photo of a {species} {individual id} in {freezing, cold, chilly, cool, warm, hot} temperature, with face direction {front, back, left, right}, captured during the {day, night}." Figure 4 shows this as "A photo of a Stoat <Bob> in <cool> temperature, with face direction <left>, captured during the <day>.""

    The ReID evaluation task is to match query and gallery images without access to the identity label, but the prompt template inserts the ground-truth {individual id} into the text input. The text branch is trained with identity classification and the cross-attention loss in Eq. (4) explicitly aligns the text embedding of this prompt with the image embedding. At inference, if the same template is used, the text encoder can directly read the answer from the prompt; the reported mAP gains (e.g., +6.5 on Penguin in Table 2) then reduce to label exposure rather than to environmental metadata. The paper never states that {individual id} is dropped, replaced, or masked at inference, so as written the prediction is defined in terms of the label it claims to predict.

full rationale

The MetaWild dataset construction and the MFA architecture itself are not inherently circular, and the paper's self-citations are not load-bearing. The circularity is concentrated in the evaluation protocol as described: the text prompt includes the ground-truth individual identity token. Since the identity token is exactly what ReID is supposed to infer, any improvement from MFA could be achieved by reading that token, making the central claim—that environmental metadata improves ReID—unsupported by the reported experiments. The remaining metadata (temperature, day/night, face orientation) is not circular, and if the authors were to confirm that the identity token is absent or anonymized at inference, the claim would have independent content. However, the paper provides no such statement or ablation, so the central result is forced by construction as written.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the MFA's fusion of metadata into CLIP embeddings. The unstated test-time prompt behavior is the largest uncertainty; temperature binning and loss temperature are hand-chosen constants. No new physical or conceptual entities are introduced.

free parameters (2)
  • Temperature category boundaries = not reported
    Numeric temperature is converted into six words (freezing, cold, chilly, cool, warm, hot) using a comfort scale from Gagge et al. [16]; the bin edges change the text prompt and are hand-chosen.
  • Softmax temperature tau in MFA loss = not reported
    Equation (4) uses a temperature parameter tau in the contrastive loss for the cross-attention module; its value is unspecified and influences the training objective.
assumptions (3)
  • domain assumption Pretrained CLIP encoders provide meaningful vision-language alignment
    Section 4.1 uses the CLIP text encoder T(·) and image encoder; the method inherits assumptions about the quality and alignment of these pretrained encoders.
  • domain assumption Identity labels from temporal windows and manual inspection are correct
    Section 3.2 groups animals captured within narrow time windows at the same camera as the same individual; the ReID evaluation depends on these labels being accurate.
  • ad hoc to paper Test-time prompts exclude the individual identity token
    The prompt template in Section 4.1 includes {individual id}; for the evaluation to be valid, this token must be absent or masked at inference, but the paper does not state this.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MetaWild: A Multimodal Dataset for Animal Re-Identification with Environmental Metadata." pith.science (2026). https://pith.science/paper/KDNLGBVF

@misc{pith2026250113368,
  author       = {Pith},
  title        = {Pith review of: MetaWild: A Multimodal Dataset for Animal Re-Identification with Environmental Metadata},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KDNLGBVF}},
  note         = {Machine review of arXiv:2501.13368}
}
read the original abstract

Identifying individual animals within large wildlife populations is essential for effective wildlife monitoring and conservation efforts. Recent advancements in computer vision have shown promise in animal re-identification (Animal ReID) by leveraging data from camera traps. However, existing Animal ReID datasets rely exclusively on visual data, overlooking environmental metadata that ecologists have identified as highly correlated with animal behavior and identity, such as temperature and circadian rhythms. Moreover, the emergence of multimodal models capable of jointly processing visual and textual data presents new opportunities for Animal ReID, but existing datasets fail to leverage these models' text-processing capabilities, limiting their full potential. Additionally, to facilitate the use of metadata in existing ReID methods, we propose the Meta-Feature Adapter (MFA), a lightweight module that can be incorporated into existing vision-language model (VLM)-based Animal ReID methods, allowing ReID models to leverage both environmental metadata and visual information to improve ReID performance. Experiments on MetaWild show that combining baseline ReID models with MFA to incorporate metadata consistently improves performance compared to using visual information alone, validating the effectiveness of incorporating metadata in re-identification. We hope that our proposed dataset can inspire further exploration of multimodal approaches for Animal ReID.

Figures

Figures reproduced from arXiv: 2501.13368 by the authors.

Figure 1
Figure 1. Overview of multimodal Animal ReID framework [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Examples of the MetaWild dataset [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Example images showing different face orientations [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Overview of the proposed Meta-Feature Adapter (MFA) module [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 6
Figure 6. Figure 6: Challenging ReID cases where incorporating meta [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 39 canonical work pages

  1. [1]

    Labeled Information Library of Alexandria: Biology and Conserva- tion (LILA)

    Accessed: 2024. Labeled Information Library of Alexandria: Biology and Conserva- tion (LILA). https://lila.science/

    work page 2024

  2. [2]

    Trail Camera Images of New Zealand Animals

    Accessed: 2024. Trail Camera Images of New Zealand Animals. https://lila.science/ datasets/nz-trailcams

    work page 2024

  3. [3]

    Community Data License Agreement – Permissive – Version 1.0

    Accessed: 2025. Community Data License Agreement – Permissive – Version 1.0. https://cdla.dev/permissive-1-0/

    work page 2025

  4. [4]

    New Zealand’s native animals

    Accessed: 2025. New Zealand’s native animals. https://www.doc.govt.nz/nature/ native-animals/

    work page 2025

  5. [5]

    New Zealand’s unique biodiversity is at risk from pests, weeds and other threats

    Accessed: 2025. New Zealand’s unique biodiversity is at risk from pests, weeds and other threats. https://www.doc.govt.nz/nature/pests-and-threats/

    work page 2025

  6. [6]

    Lukáš Adam, Vojtěch Čermák, Kostas Papafitsoros, and Lukas Picek. 2024. Sea- TurtleID2022: A long-span dataset for reliable sea turtle re-identification. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . 7146–7156

    work page 2024

  7. [7]

    Lukáš Adam, Kostas Papafitsoros, Claire Jean, ALan F Rees, and Vojtěch Čermák

    work page

  8. [8]

    Jean-Baptiste Alayrac, Jeff Donahue, Pauline Luc, Antoine Miech, Iain Barr, Yana Hasson, Karel Lenc, Arthur Mensch, Katherine Millican, Malcolm Reynolds, et al

Show all 46 references
  1. [9]

    Sara Meghan Beery. 2023. Where the Wild Things Are: Computer Vision for Global- Scale Biodiversity Monitoring. California Institute of Technology

    2023

  2. [10]

    David E Cade, William T Gough, Max F Czapanskiy, James A Fahlbusch, Shirel R Kahane-Rapport, Jacob MJ Linsky, Ross C Nichols, William K Oestreich, Danuta M Wisniewska, Ari S Friedlaender, et al. 2021. Tools for integrating inertial sensor data with video bio-loggers, including...

    2021

  3. [11]

    Vojtěch Čermák, Lukas Picek, Lukáš Adam, and Kostas Papafitsoros. 2024. WildlifeDatasets: An open-source toolkit for animal re-identification. In Pro- ceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . 5953–5963

    2024

  4. [12]

    Ted Cheeseman, Ken Southerland, Jinmo Park, Marilia Olio, Kiirsten Flynn, John Calambokidis, Lindsey Jones, Claire Garrigue, Astrid Frisch Jordán, Addison Howard, et al . 2022. Advanced image recognition: a fully automated, high- accuracy photo-identification matching system f...

    2022

  5. [13]

    Bowen Chen, Yun Sing Koh, and Gillian Dobbie. 2024. SSAT-Adapter: Enhancing Vision-Language Model Few-shot Learning with Auxiliary Tasks. In Proceedings of the 32nd ACM International Conference on Multimedia . 1004–1013

    2024

  6. [14]

    Liangyu Chen, Bo Li, Sheng Shen, Jingkang Yang, Chunyuan Li, Kurt Keutzer, Trevor Darrell, and Ziwei Liu. 2024. Large language models are visual reasoning coordinators. Advances in Neural Information Processing Systems 36 (2024)

    2024

  7. [15]

    Andrew Cossins. 2012. Temperature biology of animals . Springer Science & Business Media

    2012

  8. [16]

    A Pharo Gagge, Jan AJ Stolwijk, and James Daniel Hardy. 1967. Comfort and thermal sensations and associated physiological responses at various ambient temperatures. Environmental Research 1, 1 (1967), 1–20

    1967

  9. [17]

    Jing Gao, Tilo Burghardt, William Andrew, Andrew W Dowsey, and Neill W Campbell. 2021. Towards self-supervision for video identification of individual holstein-friesian cattle: The Cows2021 dataset. arXiv preprint arXiv:2105.01938 (2021)

    2021 arXiv

  10. [18]

    Peng Gao, Shijie Geng, Renrui Zhang, Teli Ma, Rongyao Fang, Yongfeng Zhang, Hongsheng Li, and Yu Qiao. 2024. Clip-adapter: Better vision-language models with feature adapters. International Journal of Computer Vision 132, 2 (2024), 581–595

    2024

  11. [19]

    Zhimin He, Jiangbo Qian, Diqun Yan, Chong Wang, and Yu Xin. 2023. Ani- mal re-identification algorithm for posture diversity. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing . IEEE, 1–5

    2023

  12. [20]

    Sven Heiling, Santosh Khanal, Aiko Barsch, Gabriela Zurek, Ian T Baldwin, and Emmanuel Gaquerel. 2016. Using the knowns to discover the unknowns: MS- based dereplication uncovers structural diversity in 17-hydroxygeranyllinalool diterpene glycoside production in the Solanaceae...

    2016

  13. [21]

    Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig. 2021. Scaling up visual and vision- language representation learning with noisy text supervision. In International Conference on Machine Learning . PMLR, 4904–4916

    2021

  14. [22]

    Bingliang Jiao, Lingqiao Liu, Liying Gao, Ruiqi Wu, Guosheng Lin, Peng Wang, and Yanning Zhang. 2024. Toward re-identifying any animal. Advances in Neural Information Processing Systems 36 (2024)

    2024

  15. [23]

    Daria Kern, Tobias Schiele, Ulrich Klauck, and Winfred Ingabire. 2024. Towards Automated Chicken Monitoring: Dataset and Machine Learning Methods for Visual, Noninvasive Reidentification. Animals 15, 1 (2024), 1

    2024

  16. [24]

    Lisette MC Leliveld, Elisabetta Riva, Gabriele Mattachini, Alberto Finzi, Daniela Lovarelli, and Giorgio Provolo. 2022. Dairy cow behavior is affected by period, time of day and housing. Animals 12, 4 (2022), 512

    2022

  17. [25]

    Dangwei Li, Zhang Zhang, Xiaotang Chen, and Kaiqi Huang. 2018. A richly annotated pedestrian dataset for person retrieval in real surveillance scenarios. IEEE Transactions on Image Processing 28, 4 (2018), 1575–1590

    2018

  18. [26]

    Shuyuan Li, Jianguo Li, Hanlin Tang, Rui Qian, and Weiyao Lin. 2020. ATRW: A Benchmark for Amur Tiger Re-identification in the Wild. In Proceedings of the 28th ACM International Conference on Multimedia . 2590–2598

    2020

  19. [27]

    Siyuan Li, Li Sun, and Qingli Li. 2023. CLIP-ReID: exploiting vision-language model for image re-identification without concrete text labels. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 37. 1405–1413

    2023

  20. [28]

    Catherine McVey, Fushing Hsieh, Diego Manriquez, Pablo Pinedo, and Kristina Horback. 2023. Invited Review: Applications of unsupervised machine learning in livestock behavior: Case studies in recovering unanticipated behavioral patterns from precision livestock farming data st...

    2023

  21. [29]

    Kostas Papafitsoros, Aliki Panagopoulou, and Gail Schofield. 2021. Social media reveals consistently disproportionate tourism pressure on a threatened marine vertebrate. Animal Conservation 24, 4 (2021), 568–579

    2021

  22. [30]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In International Conference on Machine Learni...

    2021

  23. [31]

    Stefan Schneider, Graham W Taylor, Stefan Linquist, and Stefan C Kremer

  24. [32]

    Gail Schofield, Kostas Papafitsoros, Chloe Chapman, Akanksha Shah, Lucy West- over, Liam CD Dickson, and Kostas A Katselidis. 2022. More aggressive sea turtles win fights over foraging resources independent of body size and years of presence. Animal Behaviour 190 (2022), 209–219

    2022

  25. [33]

    Xinyu Shi, Dong Wei, Yu Zhang, Donghuan Lu, Munan Ning, Jiashun Chen, Kai Ma, and Yefeng Zheng. 2022. Dense cross-query-and-support attention weighted mask aggregation for few-shot segmentation. In European Conference on Computer Vision. Springer, 151–168

    2022

  26. [34]

    Laurens Van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of Machine Learning Research 9, 11 (2008)

    2008

  27. [35]

    Anjli Varghese, Malathy Jawahar, and A Amalin Prince. 2023. Fine-tuning Con- vNets with novel leather image data for species identification. In Fifteenth Inter- national Conference on Machine Vision , Vol. 12701. SPIE, 150–157

    2023

  28. [36]

    Oscar Wahltinez and Sarah J Wahltinez. 2024. An open-source general purpose machine learning framework for individual animal re-identification using few- shot learning. Methods in Ecology and Evolution 15, 2 (2024), 373–387

    2024

  29. [37]

    Feng Wang and Huaping Liu. 2021. Understanding the behaviour of contrastive loss. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2495–2504

    2021

  30. [38]

    Hendrik Weideman, Chuck Stewart, Jason Parham, Jason Holmberg, Kiirsten Flynn, John Calambokidis, D Barry Paul, Anka Bedetti, Michelle Henley, Frank Pope, et al. 2020. Extracting identifying contours for African elephants and hump- back whales using a learned appearance model....

    2020

  31. [39]

    Pengfei Xu, Yuanyuan Zhang, Minghao Ji, Songtao Guo, Zhanyong Tang, Xiang Wang, Jing Guo, Junjie Zhang, and Ziyu Guan. 2024. Advanced intelligent monitoring technologies for animals: A survey. Neurocomputing 585 (2024), 127640

    2024

  32. [40]

    Han Yu, Xingxuan Zhang, Renzhe Xu, Jiashuo Liu, Yue He, and Peng Cui. 2024. Rethinking the evaluation protocol of domain generalization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 21897–21908

    2024

  33. [41]

    Jialu Zhang, Xinyi Wang, Chenglin Yao, Jianfeng Ren, and Xudong Jiang. 2024. Visual-linguistic Cross-domain Feature Learning with Group Attention and Gamma-correct Gated Fusion for Extracting Commonsense Knowledge. In Pro- ceedings of the 32nd ACM International Conference on M...

    2024

  34. [42]

    Tingting Zhang, Qijun Zhao, Cuo Da, Liyuan Zhou, Lei Li, and Suonan Jian- cuo. 2021. Yakreid-103: A benchmark for yak re-identification. In 2021 IEEE International Joint Conference on Biometrics . IEEE, 1–8

    2021

  35. [43]

    Di Zhao, Yun Sing Koh, Gillian Dobbie, Hongsheng Hu, and Philippe Fournier- Viger. 2024. Symmetric Self-Paced Learning for Domain Generalization. In Pro- ceedings of the AAAI Conference on Artificial Intelligence , Vol. 38. 16961–16969

    2024

  36. [2019]

    Methods in Ecology and Evolution 10, 4 (2019), 461–470

    Past, present and future approaches using computer vision for animal re-identification from camera trap data. Methods in Ecology and Evolution 10, 4 (2019), 461–470

    2019

  37. [2022]

    Advances in Neural Information Processing Systems 35 (2022), 23716–23736

    Flamingo: a visual language model for few-shot learning. Advances in Neural Information Processing Systems 35 (2022), 23716–23736

    2022

  38. [2025]

    Ecological Informatics (2025), 103158

    Exploiting facial side similarities to improve AI-driven sea turtle photo- identification systems. Ecological Informatics (2025), 103158

    2025

Pith tools

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