Pith. sign in

REVIEW 4 major objections 6 minor 96 references

An aerial color image anomaly dataset for search missions in complex forested terrain

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The Weitefeld dataset, built from a real forest manhunt, supplies 34,424 anomaly labels across 10,659 aerial images and shows current detectors miss most occluded clues.

desk verdict Valuable first-of-its-kind dataset for forested SAR anomaly detection, but the multi-view label generation needs validation and the benchmark text has a few errors that should be corrected before the dataset is widely used. read the letter →

arxiv 2507.15492 v1 pith:OB2ARRAH submitted 2025-07-21 cs.CV

Rakesh John Amala Arokia Nathan , Matthias Gessner , Nurullah Özkan , Marius Bock , Mohamed Youssef , Maximilian Mews , Björn Piltz , Ralf Berger
show 1 more author
Oliver Bimber
This is my paper · ORCID
classification cs.CV
keywords aerialcolorimageryanomalydetectionbenchmarkforestsearchandrescuemanhuntcrowdsourcedlabelingphotogrammetricbackprojectionoccludedobjectsboundingboxannotations
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

After a family murder in Weitefeld, Germany, a research aircraft scanned a forested search area and 160 volunteers manually combed 10,659 high-resolution aerial images for anything out of place. The paper releases the resulting dataset, Weitefeld: 34,424 bounding-box labels derived from 405 findings, each imaged many times from overlapping flight strips at roughly 4 cm per pixel, plus police ground-check protocols and 19,795 unlabeled images. The authors' central claim is that this is the first large-scale, realistic benchmark for anomaly detection in complex forested terrain for search-and-rescue and manhunt missions. Benchmarking six common color anomaly detectors on it, they find all perform poorly, especially under occlusion, and that a state-of-the-art object detector essentially fails, which motivates context-aware detection methods. A sympathetic reader should care because this turns a failed search operation into reusable evidence about what automated search tools can and cannot see in real forests.

What carries the argument

The load-bearing mechanism is photogrammetric label propagation: 405 crowd findings marked on single images are projected onto every image depicting the same object using bundle block adjustment, triangulation, and collinearity-based back-projection, exploiting up to 85-fold along-strip overlap to create 34,424 multi-view labels. Around this sits the data-generation pipeline: the MACS aerial camera with 50 MP RGB and thermal sensors, radiometric calibration and DCB de-Bayering, gamma and saturation adjustment, Reed-Xiaoli (RX) binary anomaly masks to guide volunteers, and a web frontend supporting point and bounding-box labels with classification comments. The benchmark is driven by comparing detector anomaly masks against the backprojected bounding boxes using average precision and average detection rate.

What would settle it

Have independent annotators manually re-mark a random sample of 50 of the 405 findings in every image where the object is visible, then measure pixel displacement between their bounding boxes and the photogrammetrically backprojected boxes; if a substantial fraction, say more than 10 percent, are displaced by more than a few pixels or land in the wrong image, the multi-view supervision that makes the dataset unique is unreliable.

Watch

Extended reading notes

Core claim

The central discovery is the dataset itself and what it demonstrates. Weitefeld is, according to the paper, the first anomaly-detection dataset built from an actual manhunt in densely forested terrain: 30,454 RGB images captured at 8,416 by 6,032 pixels with 3 to 5 cm ground sampling distance, of which 10,659 images in a priority zone were searched by 160 volunteers. Crowd workers reported 405 anomalies; photogrammetric bundle adjustment, triangulation, and collinearity-based back-projection expanded those findings into 34,424 labels across overlapping views, and 238 findings were checked on the ground by police. Benchmarking on this data, the paper reports that deep-learning detectors (FRE, FastFlow, EfficientAD) reach below 3.5 percent average precision, model-based detectors (RXG, RXM, PCA) reach near 100 percent detection rate but under 0.75 percent precision, and YOLOv12 object classification effectively fails with 0.016 percent average confidence. The paper's conclusion is that local, pixel-level color anomaly signals are insufficient; useful search tools must integrate broader context.

Load-bearing premise

The dataset's 34,424 labels are produced by photogrammetrically projecting 405 human findings onto every overlapping image that shows the same object, and the paper provides no accuracy assessment of that projection, so a systematic misalignment would repeat the same error across all views.

Editorial extensions

If this is right

  • None of the six tested color anomaly detectors performs well on Weitefeld: deep-learning methods stay below 3.5 percent average precision and model-based methods below 0.75 percent precision, which the paper reads as evidence that context-aware approaches are needed.
  • Automated object classification under dense vegetation is shown to be unrealistic even with a state-of-the-art detector (YOLOv12), meaning anomaly detection rather than classification is the viable automated route in such terrain.
  • Because each finding appears in multiple overlapping images, the dataset supports methods that exploit viewpoint changes and occlusion patterns, not just single-image pixel statistics.
  • The inclusion of police ground-check protocols for 238 findings turns the labels into more than pixel boxes: they carry operational relevance information that can be used to evaluate mission-oriented prioritization.
  • The unlabeled images from the two non-priority zones and the extendable web interface allow the benchmark to grow beyond the original manhunt and to support supervised fine-tuning or self-supervised pretraining.

Reading between the lines

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

  • The paper argues that image context matters but never defines or measures it; a testable next step is to build detectors that suppress anomalies recurring globally, such as tree stumps, and evaluate whether that closes the precision gap.
  • Because the photogrammetric backprojection lacks an accuracy assessment, benchmark numbers may mix detector error with label noise; measuring reprojection error on a manually re-annotated subset would tell how much of the reported failure is genuine algorithm weakness.
  • The dataset's 'anomaly' labels are human judgments of oddity rather than confirmed target locations, since the suspect was never found; recall-oriented claims should be read with that caveat, though precision-oriented evaluation is well supported by ground checks.
  • The 19,795 unlabeled images invite a natural experiment: pretrain a forest-adapted representation on them and measure whether downstream detection on the labeled priority zone improves beyond ImageNet-initialized baselines.
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. The manuscript presents Weitefeld, a new aerial RGB image dataset for anomaly detection in forested terrain, collected during a real manhunt search. It contains 10,659 labeled images with 34,424 anomaly labels produced by photogrammetrically back-projecting 405 crowd-sourced findings across overlapping views, plus 19,795 unlabeled images from adjacent areas. The authors report benchmark results for six color anomaly detectors, showing poor performance, and provide a web interface for viewing and extending the dataset.

Significance. If the multi-view labels are accurate, the dataset fills a genuine gap: no existing public dataset targets dense-forest search scenarios with centimeter-resolution imagery, heavy occlusion, and multi-view labels. The involvement of police ground verification for 238 findings is a strong practical credential, and the public release of data and code is commendable. The benchmark, while intentionally simple, gives a useful baseline of current detector limitations. The main caveat is that the backprojection step is currently unvalidated, and the label-count arithmetic contains an inconsistency; these must be resolved before the dataset can be relied upon.

major comments (4)
  1. [Data Review, Mapping, and Ground Operations; Data Records] The statement 'Of the 405 findings, 238 were flagged as relevant, while the remaining 16 ... were deemed irrelevant' is arithmetically inconsistent: 238+16=254, leaving 151 findings unaccounted for. The later claim in 'Data Records' that 'Each of the findings was identified during ground observations by police' is also contradicted by the ground-operation description, which inspected only the 238 relevant findings. Please clarify the status (relevant, irrelevant, unverified) of all 405 findings and correct the text accordingly.
  2. [Data Review, Mapping, and Ground Operations; Data Records] The 34,424 multi-view labels are the dataset's main advertised contribution, but the photogrammetric backprojection has no reported accuracy assessment. No bundle-adjustment RMSE, tie-point statistics, terrain-model specification, or comparison of backprojected boxes with manual annotations in other views is provided. Because the original labels are 2D boxes in a single image, the transfer requires a 3D localization step; for elevated objects (people, tents, hunting stands) a terrain-intersection assumption can cause systematic horizontal offsets that grow with off-nadir angle and object height, at 3-5 cm GSD potentially tens of pixels. Since the same error is reprojected into every view, this could corrupt the multi-view supervision. Please add a validation section reporting the geometric accuracy of the backprojection and a sample of manual cross-view checks. The Limitations section should also mention this residual risk.
  3. [Online Crowd Search; Technical Validation] The binary anomaly masks shown to volunteers during labeling were generated with the RX detector, and the benchmark then evaluates RX-G on labels that may be biased by exposure to those masks. While volunteers were instructed that masks are only a supplementary aid, the labeling process is not independent of the detector being benchmarked, so the reported RX-G detection rates may be partially circular. Please discuss this contamination and, if possible, quantify its effect (e.g., by comparing labels collected with and without mask viewing on a subset of images), or restrict the benchmark claim accordingly.
  4. [Technical Validation] The benchmark is restricted to the original single-view labels ('omitting backprojections'), so the 34,424 multi-view labels are neither used nor validated by the reported experiments. This is a defensible choice for a first sanity check, but the paper should state clearly that the benchmark does not exercise the multi-view portion of the dataset, and it would strengthen the data descriptor to include a small evaluation (e.g., detection rate on backprojected boxes across views) to give users confidence in the transferred labels.
minor comments (6)
  1. [Table 2] The heading 'contentious image segment' should read 'continuous image segment'.
  2. [Figure 5] The caption lists 'FIRE' but the method is 'FRE' (Feature Reconstruction Error) in the text; please unify the notation.
  3. [Related Works & Datasets] The novelty claim that no existing dataset is designed for dense-forest search should be tempered in light of NOMAD [90] and WISARD [89], which cover partially occluded aerial SAR scenarios; please discuss the specific distinction.
  4. [Technical Validation] The definition of 'average precision' as a per-image pixel-level ratio differs from the standard information-retrieval average precision; please clarify the terminology to avoid confusion with detection AP.
  5. [Downloading the Dataset] The data format description is hard to parse; a small table or formal grammar for the entries in data.txt would improve usability.
  6. [Data Records] The sentence 'Each of the findings was identified during ground observations by police' is an overstatement given that only 238 of 405 findings were ground-verified; see major comment 1.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: dataset construction and benchmark evaluation are independent of detector outputs; the only concerns are validation gaps, not circular reasoning.

full rationale

The paper is a dataset description, not a predictive derivation, and its central claims are the dataset's existence, scale, and realism. The labels were produced by 160 human volunteers who reviewed RGB images with optional assistance from RX anomaly masks, followed by manual review and police ground verification of 238 findings; the benchmark then evaluates detectors against those human-verified labels. No detector parameter is fitted to the labels, and the RX mask was a supplementary aid with an explicitly stated caveat that it could produce false positives or overlook relevant areas. The photogrammetric backprojection that expands 405 findings into 34,424 labels is unvalidated and raises a correctness concern, but it is not circular: the backprojection does not take detector outputs or benchmark results as inputs. The two self-citations [82, 83] supporting the claim that object classification fails under occlusion are not load-bearing because the paper independently reports its own YOLOv12 experiment with a 0.016% average confidence, confirming the failure within the paper itself. No self-definitional, fitted-input-called-prediction, or self-citation-chain circularity is present.

Assumptions & free parameters 4 free parameters · 2 assumptions · 0 invented entities

The dataset paper introduces no new theoretical entities. Its free parameters are hand-chosen processing choices that shape the images and the labeling aid. The main assumptions are the accuracy of the photogrammetric backprojection and the reliability of the crowd labels, both of which are load-bearing for the dataset's value.

free parameters (4)
  • RX anomaly threshold = 0.985
    Set to generate binary anomaly masks used to guide volunteers; chosen by hand, affects which areas humans inspected.
  • Gamma correction exponent = 0.37
    Applied in post-processing to adjust tonal range; hand-chosen and affects appearance.
  • Saturation scale = 2.0
    Applied to compensate for high exposure; hand-chosen and affects color distribution.
  • Exposure time = 1.3 ms
    Set to favor shaded areas, assuming the person is more likely there; a modeling choice that affects image quality.
assumptions (2)
  • domain assumption Camera georeferencing and IMU/GNSS data are accurate enough for photogrammetric backprojection of labels.
    The multi-view label propagation depends on this; no accuracy assessment is provided.
  • domain assumption Human crowd labels are correct and the police ground check validates a subset (238 findings).
    Labels are the ground truth for benchmarking; volunteer labeling is subjective and only a subset was ground-verified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An aerial color image anomaly dataset for search missions in complex forested terrain." pith.science (2026). https://pith.science/paper/OB2ARRAH

@misc{pith2026250715492,
  author       = {Pith},
  title        = {Pith review of: An aerial color image anomaly dataset for search missions in complex forested terrain},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OB2ARRAH}},
  note         = {Machine review of arXiv:2507.15492}
}
read the original abstract

After a family murder in rural Germany, authorities failed to locate the suspect in a vast forest despite a massive search. To aid the search, a research aircraft captured high-resolution aerial imagery. Due to dense vegetation obscuring small clues, automated analysis was ineffective, prompting a crowd-search initiative. This effort produced a unique dataset of labeled, hard-to-detect anomalies under occluded, real-world conditions. It can serve as a benchmark for improving anomaly detection approaches in complex forest environments, supporting manhunts and rescue operations. Initial benchmark tests showed existing methods performed poorly, highlighting the need for context-aware approaches. The dataset is openly accessible for offline processing. An additional interactive web interface supports online viewing and dynamic growth by allowing users to annotate and submit new findings.

Figures

Figures reproduced from arXiv: 2507.15492 by the authors.

Figure 1
Figure 1. Map of scan section and priority zone with orthographically projected aerial images of the flight and locations of relevant findings (top). Example finding in full aerial image and close-up (bottom). During ground observation, this finding was later identified as "a small barrel, very old, partially overgrown and filled with soil". caused by vegetation. A key reason for this failure is that decisions based on local … view at source ↗
Figure 2
Figure 2. Close-ups of various sample findings contained in the dataset together with their bounding-box labels. Overall, they include objects such as barrels, trash bags, tarps, metal barriers, floating objects on water; but also people and potential man-made shelters, hunting stands, sheds, huts, tents, shooting ranges, fire pits, and others – all identified based on color and structural anomalies. of samples, enabling eval… view at source ↗
Figure 3
Figure 3. The research aircraft (a Stemme S10 motorglider, top-left) was equipped with the Modular Aerial Camera System (MACS, bottom-left) mounted in one of its underwing pods. This system was used for data acquisition. The flight plan (right) displays the locations and numbering of the 15 flight strips. GPS coordinates for Weitefeld: 50.72468°N (latitude), 7.92714°E (longitude). Methods The following subsection summarizes t… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Image analyzer web-frontend used for online crowd search and subsequent reviews of findings. It supports browsing image sequences from the same flight strip, toggling between RGB images and anomaly masks, zooming into specific areas, panning across details, adjusting b…
Figure 5
Figure 5. Figure 5: A comparison of common deep-learning-based (FIRE, FastFlow, EfficientAD) and model-based (RX-G, RX-M, PCA) color anomaly detectors, evaluated in terms of average precision (top) and average detection rate (bottom) across a range of anomaly thresholds t (x-axis). 10/17 …
Figure 6
Figure 6. Figure 6: Performance of deep-learning-based color anomaly detectors (FRE, FastFlow, EfficientAD) over each class (unknown, shelter, object, person). Anomaly threshold vs. average precision (left) and vs. average detection rate (right) [PITH_FULL_IMAGE:figures/full_fig_p017_6.png]
Figure 7
Figure 7. Figure 7: Performance of model-based color anomaly detectors (RXG,RXM,PCA) over each class (unknown, shelter, object, person). Anomaly threshold vs. average precision (left) and vs. average detection rate (right). 17/17 [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

96 extracted references · 74 canonical work pages

  1. [1]

    Adaptive multiple-band CFAR detection of an optical pattern with unknown spectral distribution

    Irving S Reed and Xiaoli Yu. “Adaptive multiple-band CFAR detection of an optical pattern with unknown spectral distribution”. In: IEEE transactions on acoustics, speech, and signal processing 38.10 (1990), pp. 1760–1770. 11/17

    1990

  2. [2]

    FRE: a fast method for anomaly detection and segmentation

    Ibrahima Ndiour et al. “FRE: a fast method for anomaly detection and segmentation”. In:arXiv preprint arXiv:2211.12650 (2022)

    arXiv 2022

  3. [3]

    Fastflow: Unsupervised anomaly detection and localization via 2d normalizing flows

    Jiawei Yu et al. “Fastflow: Unsupervised anomaly detection and localization via 2d normalizing flows”. In:arXiv preprint arXiv:2111.07677 (2021)

    arXiv 2021

  4. [4]

    Efficientad: Accurate visual anomaly detection at millisecond-level latencies

    Kilian Batzner, Lars Heckler, and Rebecca König. “Efficientad: Accurate visual anomaly detection at millisecond-level latencies”. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2024, pp. 128–138

    2024

  5. [5]

    Anomaly detection and classification for hyperspectral imagery

    Chein-I Chang and Shao-Shan Chiang. “Anomaly detection and classification for hyperspectral imagery”. In: IEEE transactions on geoscience and remote sensing 40.6 (2002), pp. 1314–1325

    2002

  6. [6]

    A novel anomaly detection scheme based on principal component classifier

    Mei-Ling Shyu et al. “A novel anomaly detection scheme based on principal component classifier”. In: Proceedings of the IEEE foundations and new directions of data mining workshop. IEEE Press Piscataway, NJ, USA. 2003, pp. 172–179

    2003

  7. [7]

    Spectral anomaly detection with machine learning for wilderness search and rescue

    Julia Proft, Jesus Suarez, and Robin Murphy. “Spectral anomaly detection with machine learning for wilderness search and rescue”. In: 2015 IEEE MIT Undergraduate Research Technology Conference (URTC). IEEE. 2015, pp. 1–3

    2015

  8. [8]

    Color anomaly detection and suggestion for wilderness search and rescue

    Bryan S Morse, Daniel Thornton, and Michael A Goodrich. “Color anomaly detection and suggestion for wilderness search and rescue”. In: Proceedings of the seventh annual ACM/IEEE international conference on Human-Robot Interaction. 2012, pp. 455–462

    2012

Show all 96 references
  1. [9]

    Color outlier detection for search and rescue

    Travis Marshall and L Nathan Perkins. “Color outlier detection for search and rescue”. In: Dept. Elect. Comput. Eng., Boston Univ., Boston, MA, USA, Tech. Rep. ECE-2015-01(2015)

    2015

  2. [10]

    Anomaly color detection on uav images for search and rescue works

    Dao Khanh Hoai and Nguyen Van Phuong. “Anomaly color detection on uav images for search and rescue works”. In: 2017 9th International Conference on Knowledge and Systems Engineering (KSE). IEEE. 2017, pp. 287–291

    2017

  3. [11]

    A DWT-SVDD Anomaly Detection Algorithm Using Hyperspectral Remote Sensing in Marine Search and Rescue

    Can Cui and Jin Xu. “A DWT-SVDD Anomaly Detection Algorithm Using Hyperspectral Remote Sensing in Marine Search and Rescue”. In: 2021 IEEE 6th International Conference on Signal and Image Processing (ICSIP). IEEE. 2021, pp. 181–184

    2021

  4. [12]

    ResQuad: Toward a semi-autonomous wilderness search and rescue unmanned aerial system

    M Talha Agcayazi et al. “ResQuad: Toward a semi-autonomous wilderness search and rescue unmanned aerial system”. In: 2016 International Conference on Unmanned Aircraft Systems (ICUAS). IEEE. 2016, pp. 898–904

    2016

  5. [13]

    Anomaly Detection from Aerial Images for Search and Rescue Missions

    Luca Morandini. “Anomaly Detection from Aerial Images for Search and Rescue Missions”. Available at POLITesi repository. M.Sc. Thesis, School of Industrial and Information Engineering. Politecnico di Milano, May 2021. URL: https://hdl.handle.net/10589/206674

    2021

  6. [14]

    Anomaly Detection Methods for Maritime Search and Rescue

    Ryan Sime and Rohan Loveland. “Anomaly Detection Methods for Maritime Search and Rescue”. In: Proceedings of the 14th International Conference on Pattern Recognition Applications and Methods (ICPRAM). V ol. 1. INSTICC. SciTePress, 2025, pp. 764–770. ISBN : 978-989-758-730-6. D...

    2025 doi

  7. [15]

    Evaluation of color anomaly detection in multispectral images for synthetic aperture sensing

    Francis Seits, Indrajit Kurmi, and Oliver Bimber. “Evaluation of color anomaly detection in multispectral images for synthetic aperture sensing”. In: Eng 3.4 (2022), pp. 541–553

    2022

  8. [16]

    Anomaly detection in aerial imagery using color and texture features

    Fabian Zavala-Vazquez et al. “Anomaly detection in aerial imagery using color and texture features”. In: 2019 In- ternational Conference on Mechatronics, Electronics and Automotive Engineering (ICMEAE). IEEE. 2019, pp. 45– 49

    2019

  9. [17]

    Anomaly Detection in Species Distribution Patterns: A Spatio-Temporal Approach for Biodiversity Conservation

    Mingyang He and Hao Chen. “Anomaly Detection in Species Distribution Patterns: A Spatio-Temporal Approach for Biodiversity Conservation”. In: Journal of Biobased Materials and Bioenergy 18.1 (2024), pp. 39–50

    2024

  10. [18]

    Anomaly detection for the automated visual inspection of pet preform closures

    Oliver Rippel et al. “Anomaly detection for the automated visual inspection of pet preform closures”. In: 2021 26th IEEE international conference on emerging technologies and factory automation (ETFA). IEEE. 2021, pp. 1–7

    2021

  11. [19]

    Developing an Anomaly Detection System for Automatic Defective Products’ Inspection

    Yu-Hsin Hung. “Developing an Anomaly Detection System for Automatic Defective Products’ Inspection”. In:Processes 10.8 (2022), p. 1476

    2022

  12. [20]

    A Collaborative Robotic Approach for Inspection and Anomaly Detection in Industrial Applica- tions

    Miaolong Yuan et al. “A Collaborative Robotic Approach for Inspection and Anomaly Detection in Industrial Applica- tions”. In: Social Robotics: 13th International Conference, ICSR 2021, Singapore, Singapore, November 10–13, 2021, Proceedings 13. Springer. 2021, pp. 757–762

    2021

  13. [21]

    Anomaly detection for industrial surface inspection: application in maintenance of aircraft components

    Falko Kähler, Ole Schmedemann, and Thorsten Schüppstuhl. “Anomaly detection for industrial surface inspection: application in maintenance of aircraft components”. In: Procedia CIRP 107 (2022), pp. 246–251

    2022

  14. [22]

    Industrial anomaly detection and one-class classification using generative adversarial networks

    Yu-Ting K Lai et al. “Industrial anomaly detection and one-class classification using generative adversarial networks”. In: 2018 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM). IEEE. 2018, pp. 1444–1449. 12/17

    2018

  15. [23]

    Automated optical inspection using anomaly detection and unsupervised defect clustering

    Jan Lehr et al. “Automated optical inspection using anomaly detection and unsupervised defect clustering”. In:2020 25th ieee international conference on emerging technologies and factory automation (etfa). V ol. 1. IEEE. 2020, pp. 1235– 1238

    2020

  16. [24]

    Anomaly detection using deep learning based image completion

    Matthias Haselmann, Dieter P Gruber, and Paul Tabatabai. “Anomaly detection using deep learning based image completion”. In: 2018 17th IEEE international conference on machine learning and applications (ICMLA). IEEE. 2018, pp. 1237–1242

    2018

  17. [25]

    Anomaly detection neural network with dual auto-encoders GAN and its industrial inspection applications

    Ta-Wei Tang et al. “Anomaly detection neural network with dual auto-encoders GAN and its industrial inspection applications”. In: Sensors 20.12 (2020), p. 3336

    2020

  18. [26]

    Anomaly Detection for Industrial Inspection using Convolutional Autoencoder and Deep Feature-based One-class Classification

    Jamal Saeedi and Alessandro Giusti. “Anomaly Detection for Industrial Inspection using Convolutional Autoencoder and Deep Feature-based One-class Classification.” In: VISIGRAPP (5: VISAPP). 2022, pp. 85–96

    2022

  19. [27]

    Advanced Color Pseudo Anomaly to Enhance Learning of Convolutional Neural Network Models

    Rizwan Ali Shah and HyungWon Kim. “Advanced Color Pseudo Anomaly to Enhance Learning of Convolutional Neural Network Models”. In: 2023 IEEE International Conference on Consumer Electronics-Asia (ICCE-Asia). IEEE. 2023, pp. 1–3

    2023

  20. [28]

    Anomaly detection in fundus images by self-adaptive decomposition via local and color based sparse coding

    Yuchen Du et al. “Anomaly detection in fundus images by self-adaptive decomposition via local and color based sparse coding”. In: Biomedical Optics Express 13.8 (2022), pp. 4261–4277

    2022

  21. [29]

    Unsupervised deep learning approach for structural anomaly detection using probabilistic features

    Hua-Ping Wan et al. “Unsupervised deep learning approach for structural anomaly detection using probabilistic features”. In: Structural Health Monitoring 24.1 (2025), pp. 3–33

    2025

  22. [30]

    Beyond dents and scratches: Logical constraints in unsupervised anomaly detection and localiza- tion

    Paul Bergmann et al. “Beyond dents and scratches: Logical constraints in unsupervised anomaly detection and localiza- tion”. In: International Journal of Computer Vision 130.4 (2022), pp. 947–969

    2022

  23. [31]

    Anomaly detection in thermal images using deep neural networks

    Cai Lile and Li Yiqun. “Anomaly detection in thermal images using deep neural networks”. In: 2017 IEEE International conference on image processing (ICIP). IEEE. 2017, pp. 2299–2303

    2017

  24. [32]

    Thermal anomaly detection based on saliency analysis from multimodal imaging sources

    A Sledz and C Heipke. “Thermal anomaly detection based on saliency analysis from multimodal imaging sources”. In: (2021)

    2021

  25. [33]

    Model-based thermal anomaly detection in cloud datacenters using thermal imaging

    Eun Kyung Lee, Hariharasudhan Viswanathan, and Dario Pompili. “Model-based thermal anomaly detection in cloud datacenters using thermal imaging”. In: IEEE Transactions on Cloud Computing 6.2 (2015), pp. 330–343

    2015

  26. [34]

    UA V-based thermal anomaly detection for distributed heating networks

    A Sledz, J Unger, and C Heipke. “UA V-based thermal anomaly detection for distributed heating networks”. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences43 (2020), pp. 499–505

    2020

  27. [35]

    Towards detection of thermal anomalies in large urban areas using simulation

    Eva Burkard, Dimitri Bulatov, and Benedikt Kottler. “Towards detection of thermal anomalies in large urban areas using simulation”. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences 43 (2020), pp. 1195–1202

    2020

  28. [36]

    Seasons in drift: A long-term thermal imaging dataset for studying concept drift

    Ivan Adriyanov Nikolov et al. “Seasons in drift: A long-term thermal imaging dataset for studying concept drift”. In: Thirty-fifth Conference on Neural Information Processing Systems. Neural Information Processing Systems Foundation. 2021

    2021

  29. [37]

    Anomaly Detection for Hotspot Identification in Landsat Images

    Antoine Tadros et al. “Anomaly Detection for Hotspot Identification in Landsat Images”. In: IGARSS 2024-2024 IEEE International Geoscience and Remote Sensing Symposium. IEEE. 2024, pp. 8218–8222

    2024

  30. [38]

    Detecting district heating leaks in thermal imagery: Comparison of anomaly detection methods

    Elena V ollmer et al. “Detecting district heating leaks in thermal imagery: Comparison of anomaly detection methods”. In: Automation in Construction 168 (2024), p. 105709

    2024

  31. [39]

    Hyperspectral anomaly detection with multiscale attribute and edge-preserving filters

    Shutao Li et al. “Hyperspectral anomaly detection with multiscale attribute and edge-preserving filters”. In: IEEE Geoscience and Remote Sensing Letters 15.10 (2018), pp. 1605–1609

    2018

  32. [40]

    Data model LUT-based change and anomaly detection for real-time multispectral image characterization

    Holger Jaenisch and James Handley. “Data model LUT-based change and anomaly detection for real-time multispectral image characterization”. In: Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XVIII. V ol. 8390. SPIE. 2012, pp. 323–335

    2012

  33. [41]

    Anomaly detection and classification in multispectral time series based on hidden Markov models

    Kareth M Leon-Lopez et al. “Anomaly detection and classification in multispectral time series based on hidden Markov models”. In: IEEE transactions on geoscience and remote sensing 60 (2021), pp. 1–11

    2021

  34. [42]

    Temporal anomaly detection in multispectral imagery

    Amanda Ziemann, Hope Simonoko, and Eric Flynn. “Temporal anomaly detection in multispectral imagery”. In:IGARSS 2020-2020 IEEE International Geoscience and Remote Sensing Symposium. IEEE. 2020, pp. 3975–3978

    2020

  35. [43]

    Hyperspectral anomaly detection: A survey

    Hongjun Su et al. “Hyperspectral anomaly detection: A survey”. In: IEEE Geoscience and Remote Sensing Magazine 10.1 (2021), pp. 64–90. 13/17

    2021

  36. [44]

    Hyperspectral anomaly detection based on variational background inference and generative adversarial network

    Zhiwei Wang et al. “Hyperspectral anomaly detection based on variational background inference and generative adversarial network”. In: Pattern Recognition 143 (2023), p. 109795

    2023

  37. [45]

    Hyperspectral anomaly detection via low-rank and sparse decomposition with cluster subspace accumulation

    Baozhi Cheng and Yan Gao. “Hyperspectral anomaly detection via low-rank and sparse decomposition with cluster subspace accumulation”. In: Scientific Reports 14.1 (2024), p. 29530

    2024

  38. [46]

    Hyperspectral anomaly detection: a performance comparison of existing techniques

    Noman Raza Shah et al. “Hyperspectral anomaly detection: a performance comparison of existing techniques”. In: International Journal of Digital Earth 15.1 (2022), pp. 2078–2125

    2022

  39. [47]

    Anomaly detection in hyperspectral images based on low-rank and sparse representation

    Yang Xu et al. “Anomaly detection in hyperspectral images based on low-rank and sparse representation”. In: IEEE Transactions on Geoscience and Remote Sensing 54.4 (2015), pp. 1990–2000

    2015

  40. [48]

    Spectral–spatial feature extraction for hyperspectral anomaly detection

    Jie Lei et al. “Spectral–spatial feature extraction for hyperspectral anomaly detection”. In: IEEE Transactions on Geoscience and Remote Sensing 57.10 (2019), pp. 8131–8143

    2019

  41. [49]

    Deep industrial image anomaly detection: A survey

    Jiaqi Liu et al. “Deep industrial image anomaly detection: A survey”. In: Machine Intelligence Research 21.1 (2024), pp. 104–135

    2024

  42. [50]

    Hyperspectral anomaly detection using deep learning: A review

    Xing Hu et al. “Hyperspectral anomaly detection using deep learning: A review”. In: Remote Sensing 14.9 (2022), p. 1973

    2022

  43. [51]

    A deep learning based technique for anomaly detection in surveillance videos

    Prakhar Singh and Vinod Pankajakshan. “A deep learning based technique for anomaly detection in surveillance videos”. In: 2018 Twenty Fourth National Conference on Communications (NCC). IEEE. 2018, pp. 1–6

    2018

  44. [52]

    DeepAnT: A deep learning approach for unsupervised anomaly detection in time series

    Mohsin Munir et al. “DeepAnT: A deep learning approach for unsupervised anomaly detection in time series”. In: Ieee Access 7 (2018), pp. 1991–2005

    2018

  45. [53]

    Deep learning for anomaly detection

    Ruoying Wang et al. “Deep learning for anomaly detection”. In:Proceedings of the 13th international conference on web search and data mining. 2020, pp. 894–896

    2020

  46. [54]

    A model-based anomaly detection approach for analyzing streaming aircraft engine measurement data

    Donald L Simon and Aidan W Rinehart. A model-based anomaly detection approach for analyzing streaming aircraft engine measurement data. V ol. 45752. American Society of Mechanical Engineers, 2014

    2014

  47. [55]

    Model based anomaly detection in high dimensional DATA

    Imen Gatfaoui, Basel Solaiman, and Imed Riadh Farah. “Model based anomaly detection in high dimensional DATA”. In: 2022 6th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP). IEEE. 2022, pp. 1–6

    2022

  48. [56]

    Model-Based Anomaly Detection Approaches

    Kishan G. Mehrotra, Chilukuri K. Mohan, and HuaMing Huang. “Model-Based Anomaly Detection Approaches”. In: Anomaly Detection Principles and Algorithms. Cham: Springer International Publishing, 2017, pp. 57–94. ISBN : 978-3-319-67526-8. DOI: 10.1007/978-3-319-67526-8_5 . URL: h...

    2017 doi

  49. [57]

    GAN-based anomaly detection: A review

    Xuan Xia et al. “GAN-based anomaly detection: A review”. In: Neurocomputing 493 (2022), pp. 497–535

    2022

  50. [58]

    MVTec AD–A comprehensive real-world dataset for unsupervised anomaly detection

    Paul Bergmann et al. “MVTec AD–A comprehensive real-world dataset for unsupervised anomaly detection”. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019, pp. 9592–9600

    2019

  51. [59]

    The MVTec anomaly detection dataset: a comprehensive real-world dataset for unsupervised anomaly detection

    Paul Bergmann et al. “The MVTec anomaly detection dataset: a comprehensive real-world dataset for unsupervised anomaly detection”. In: International Journal of Computer Vision 129.4 (2021), pp. 1038–1059

    2021

  52. [60]

    The mvtec ad 2 dataset: Advanced scenarios for unsupervised anomaly detection

    Lars Heckler-Kram et al. “The mvtec ad 2 dataset: Advanced scenarios for unsupervised anomaly detection”. In: arXiv preprint arXiv:2503.21622 (2025)

    2025

  53. [61]

    The mvtec 3d-ad dataset for unsupervised 3d anomaly detection and localization

    Paul Bergmann et al. “The mvtec 3d-ad dataset for unsupervised 3d anomaly detection and localization”. In: arXiv preprint arXiv:2112.09045 (2021)

    2021 arXiv

  54. [62]

    VT-ADL: A vision transformer network for image anomaly detection and localization

    Pankaj Mishra et al. “VT-ADL: A vision transformer network for image anomaly detection and localization”. In:2021 IEEE 30th International Symposium on Industrial Electronics (ISIE). IEEE. 2021, pp. 01–06

    2021

  55. [63]

    Anomaly detection in crowd scene

    Shu Wang and Zhenjiang Miao. “Anomaly detection in crowd scene”. In: IEEE 10th International Conference on Signal Processing Proceedings. IEEE. 2010, pp. 1220–1223

    2010

  56. [64]

    Abnormal event detection at 150 fps in matlab

    Cewu Lu, Jianping Shi, and Jiaya Jia. “Abnormal event detection at 150 fps in matlab”. In: Proceedings of the IEEE international conference on computer vision. 2013, pp. 2720–2727

    2013

  57. [65]

    A day on campus-an anomaly detection dataset for events in a single camera

    Mantini Pranav, Li Zhenggang, et al. “A day on campus-an anomaly detection dataset for events in a single camera”. In: Proceedings of the Asian Conference on Computer Vision. 2020

    2020

  58. [66]

    Future frame prediction for anomaly detection–a new baseline

    Wen Liu et al. “Future frame prediction for anomaly detection–a new baseline”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2018, pp. 6536–6545. 14/17

    2018

  59. [67]

    Anomaly detection in aerial videos with transformers

    Pu Jin et al. “Anomaly detection in aerial videos with transformers”. In: IEEE Transactions on Geoscience and Remote Sensing 60 (2022), pp. 1–13

    2022

  60. [68]

    Contextual Information Based Anomaly Detection for a Multi-Scene UA V Aerial Videos

    Ujjwal Verma, Radhika M Pai, et al. “Contextual Information Based Anomaly Detection for a Multi-Scene UA V Aerial Videos”. In: arXiv preprint arXiv:2203.15437 (2022)

    2022 arXiv

  61. [69]

    UIT-ADrone: A novel drone dataset for traffic anomaly detection

    Tung Minh Tran et al. “UIT-ADrone: A novel drone dataset for traffic anomaly detection”. In:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 16 (2023), pp. 5590–5601

    2023

  62. [70]

    Self-supervised masked convolutional transformer block for anomaly detection

    Neelu Madan et al. “Self-supervised masked convolutional transformer block for anomaly detection”. In: IEEE Transac- tions on Pattern Analysis and Machine Intelligence 46.1 (2023), pp. 525–542

    2023

  63. [71]

    Segmentmeifyoucan: A benchmark for anomaly segmentation

    Robin Chan et al. “Segmentmeifyoucan: A benchmark for anomaly segmentation”. In: arXiv preprint arXiv:2104.14812 (2021)

    2021 arXiv

  64. [72]

    Anomalous motion detection on highway using deep learning

    Harpreet Singh, Emily M Hand, and Kostas Alexis. “Anomalous motion detection on highway using deep learning”. In: 2020 IEEE international conference on image processing (ICIP). IEEE. 2020, pp. 1901–1905

    2020

  65. [73]

    Fishyscapes: A benchmark for safe semantic segmentation in autonomous driving

    Hermann Blum et al. “Fishyscapes: A benchmark for safe semantic segmentation in autonomous driving”. In:proceedings of the IEEE/CVF international conference on computer vision workshops. 2019, pp. 0–0

    2019

  66. [74]

    Lost and found: detecting small road hazards for self-driving vehicles

    Peter Pinggera et al. “Lost and found: detecting small road hazards for self-driving vehicles”. In: 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE. 2016, pp. 1099–1106

    2016

  67. [75]

    Spot-the-difference self-supervised pre-training for anomaly detection and segmentation

    Yang Zou et al. “Spot-the-difference self-supervised pre-training for anomaly detection and segmentation”. In: European Conference on Computer Vision. Springer. 2022, pp. 392–408

    2022

  68. [76]

    Street scene: A new dataset and evaluation protocol for video anomaly detection

    Bharathkumar Ramachandra and Michael Jones. “Street scene: A new dataset and evaluation protocol for video anomaly detection”. In: Proceedings of the IEEE/CVF winter conference on applications of computer vision. 2020, pp. 2569–2578

    2020

  69. [77]

    Spectral and spatial complexity-based hyperspectral unmixing

    Sen Jia and Yuntao Qian. “Spectral and spatial complexity-based hyperspectral unmixing”. In:IEEE Transactions on Geoscience and Remote Sensing 45.12 (2007), pp. 3867–3879

    2007

  70. [78]

    A robust nonlinear hyperspectral anomaly detection approach

    Rui Zhao, Bo Du, and Liangpei Zhang. “A robust nonlinear hyperspectral anomaly detection approach”. In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 7.4 (2014), pp. 1227–1234

    2014

  71. [79]

    Using endmembers as a coordinate system in hyperspectral imagery

    David Gillis, Jeffrey H Bowles, and Michael E Winter. “Using endmembers as a coordinate system in hyperspectral imagery”. In: Imaging Spectrometry VIII. V ol. 4816. SPIE. 2002, pp. 346–354

    2002

  72. [80]

    MUUFL Gulfport hyperspectral and LiDAR airborne data set

    Paul Gader et al. “MUUFL Gulfport hyperspectral and LiDAR airborne data set”. In: Univ. Florida, Gainesville, FL, USA, Tech. Rep. REP-2013-570 (2013)

    2013

  73. [81]

    Hyperspectral airborne “Viareggio 2013 Trial

    Nicola Acito et al. “Hyperspectral airborne “Viareggio 2013 Trial” data collection for detection algorithm assessment”. In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 9.6 (2016), pp. 2365–2376

    2016

  74. [82]

    Search and rescue with airborne optical sectioning

    David C Schedl, Indrajit Kurmi, and Oliver Bimber. “Search and rescue with airborne optical sectioning”. In:Nature Machine Intelligence 2.12 (2020), pp. 783–790

    2020

  75. [83]

    An autonomous drone for search and rescue in forests using airborne optical sectioning

    David C Schedl, Indrajit Kurmi, and Oliver Bimber. “An autonomous drone for search and rescue in forests using airborne optical sectioning”. In: Science Robotics 6.55 (2021), eabg1188

    2021

  76. [84]

    An infrared dataset for partially occluded person detection in complex environment for search and rescue

    Zhuoyuan Song et al. “An infrared dataset for partially occluded person detection in complex environment for search and rescue”. In: Scientific Data 12.1 (2025), p. 300

    2025

  77. [85]

    BIRDSAI: A dataset for detection and tracking in aerial thermal infrared videos

    Elizabeth Bondi et al. “BIRDSAI: A dataset for detection and tracking in aerial thermal infrared videos”. In:Proceedings of the IEEE/CVF Winter conference on applications of computer vision. 2020, pp. 1747–1756

    2020

  78. [86]

    SaRNet: A dataset for deep learning assisted search and rescue with satellite imagery

    Michael Thoreau and Frazer Wilson. “SaRNet: A dataset for deep learning assisted search and rescue with satellite imagery”. In: 2021 12th International Symposium on Image and Signal Processing and Analysis (ISPA). IEEE. 2021, pp. 204–208

    2021

  79. [87]

    Automatic person detection in search and rescue operations using deep CNN detectors

    Sasa Sambolek and Marina Ivasic-Kos. “Automatic person detection in search and rescue operations using deep CNN detectors”. In: Ieee Access 9 (2021), pp. 37905–37922

    2021

  80. [88]

    Deep learning approach in aerial imagery for supporting land search and rescue missions

    Dunja Boži´c-Štuli´c, Željko Maruši´c, and Sven Gotovac. “Deep learning approach in aerial imagery for supporting land search and rescue missions”. In: International Journal of Computer Vision 127.9 (2019), pp. 1256–1278

    2019

  81. [89]

    Wisard: A labeled visual and thermal image dataset for wilderness search and rescue

    Daniel Broyles, Christopher R Hayner, and Karen Leung. “Wisard: A labeled visual and thermal image dataset for wilderness search and rescue”. In: 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE. 2022, pp. 9467–9474. 15/17

    2022

  82. [90]

    NOMAD: A Natural, Occluded, Multi-scale Aerial Dataset, for Emergency Response Scenarios

    Arturo Miguel Russell Bernal, Walter Scheirer, and Jane Cleland-Huang. “NOMAD: A Natural, Occluded, Multi-scale Aerial Dataset, for Emergency Response Scenarios”. In:2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV). IEEE. 2024, pp. 8569–8580

    2024

  83. [91]

    Aerial Person Detection for Search and Rescue: Survey and Benchmarks

    Xiangqing Zhang et al. “Aerial Person Detection for Search and Rescue: Survey and Benchmarks”. In:Journal of Remote Sensing 2025 (2025), p. 0474

    2025

  84. [92]

    Drone-based RGBT tiny person detection

    Yan Zhang et al. “Drone-based RGBT tiny person detection”. In:ISPRS Journal of Photogrammetry and Remote Sensing 204 (2023), pp. 61–76

    2023

  85. [93]

    Manipal-UA V person detection dataset: A step towards benchmarking dataset and algorithms for small object detection

    KR Akshatha et al. “Manipal-UA V person detection dataset: A step towards benchmarking dataset and algorithms for small object detection”. In: ISPRS Journal of Photogrammetry and Remote Sensing 195 (2023), pp. 77–89

    2023

  86. [94]

    Capabilities and Applications of MACS Aerial Camera Systems for Environmental Research

    Jörg Brauchle et al. “Capabilities and Applications of MACS Aerial Camera Systems for Environmental Research”. In: EGU General Assembly 2021. Apr. 2021. URL: https://elib.dlr.de/142914/

    2021

  87. [95]

    Demosaicing images from colour cameras for digital image correlation

    A Forsey and S Gungor. “Demosaicing images from colour cameras for digital image correlation”. In: Optics and lasers in engineering 86 (2016), pp. 20–28

    2016

  88. [96]

    YOLOv12: Attention-Centric Real-Time Object Detectors

    Yunjie Tian, Qixiang Ye, and David Doermann. “YOLOv12: Attention-Centric Real-Time Object Detectors”. In:arXiv preprint arXiv:2502.12524 (2025). 16/17 Appendix Figure 6. Performance of deep-learning-based color anomaly detectors (FRE, FastFlow, EfficientAD) over each class (un...

    2025 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.