Pith. sign in

REVIEW 2 major objections 4 minor 40 references

Does YOLO26 Truly Offer Advantages Over Its Predecessors for Edge Deployment? A Benchmark Study in Aquaculture

T0 review · 2 major / 4 minor · reviewed 2026-07-14 · grok-4.5

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

Pith's one-line read Architectural novelty alone does not pick the best YOLO for edge aquaculture: accuracy converges, data and hardware decide.

desk verdict Solid, useful edge-deployment benchmark of YOLO26 on real RAS mortality data; architecture alone is not the story, and the numbers back that up. read the letter →

arxiv 2607.09835 v1 pith:MKMCOOR3 submitted 2026-07-10 cs.CV

Rakesh Ranjan , Gajanan S. Kothawade , Kata Sharrer , Scott Tsukuda , Christopher Good This is my paper · ORCID
classification cs.CV
keywords YOLOfishmortalitydetectionedgedeploymentprecisionaquaculturerecirculatingsystemsNMS-freeinferencelearningcurveRaspberryPi
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

This paper asks whether the newest YOLO26 detector, built for NMS-free end-to-end inference on CPU edge devices, actually beats its Ultralytics predecessors for a practical aquaculture task: spotting dead versus live fish in a recirculating tank. Twelve model variants (nano, small, medium across four generations) were trained and tested on the same mortality images, then timed on both a high-end GPU and a Raspberry Pi 5. On the full dataset every model lands within about one percentage point of mAP50, so generation barely moves final accuracy once enough labeled images exist. The real differences appear elsewhere: YOLOv8 reaches 90 percent mAP50 with only 400 images while YOLO26 nano and small need roughly 1,000, yet YOLO26n is the fastest on the Pi (7.51 FPS) and YOLOv5mu wins the medium-tier CPU race. The authors therefore argue that farms should choose models by joint constraints of data volume, target hardware, and required frame rate rather than by release date.

What carries the argument

A controlled twelve-model, seven-dataset-size learning-curve benchmark that measures mAP50, training time, and end-to-end latency on both an A100 GPU and an ONNX Runtime Raspberry Pi 5, isolating the effect of NMS-free one-to-one assignment versus earlier NMS-based heads under identical aquaculture data.

What would settle it

Repeat the identical training-size and Raspberry Pi ONNX benchmark on a multi-tank, multi-species, multi-life-stage mortality dataset that includes natural high-mortality outbreaks; if YOLO26 nano/small then match or beat YOLOv8 at 400 images, or if medium-tier CPU rankings reverse, the joint-selection claim collapses for those conditions.

Watch

Extended reading notes

Core claim

When twelve YOLO variants spanning four generations and three size tiers are trained on the same full fish-mortality dataset, final detection accuracy differs by only 1.04 percentage points of mAP50; architectural generation therefore exerts little influence on accuracy once training data are plentiful. Meaningful trade-offs instead appear in data efficiency (YOLOv8 reaches 90 percent mAP50 with 400 images; YOLO26 nano/small need 1,000) and in CPU edge speed (YOLO26n leads the nano tier at 7.51 FPS on a Raspberry Pi 5; YOLOv5mu leads the medium tier).

Load-bearing premise

The accuracy and speed rankings measured on one Atlantic-salmon grow-out tank, under two lighting regimes, with high-mortality scenes created by adding euthanized culls on the final day, are assumed representative enough of real commercial mortality and other species or life stages for the selection guidance to transfer.

Editorial extensions

If this is right

  • Farms with limited labeled mortality images should prefer YOLOv8 over YOLO26 for faster convergence to operational accuracy.
  • CPU-only edge boxes monitoring tanks can adopt YOLO26n for the highest nano-tier frame rate without sacrificing final mAP50 once data are abundant.
  • GPU-only ranking tables cannot be trusted as proxies for Raspberry-Pi or other ARM-CPU deployment choices; target-hardware timing is required.
  • Annotation effort beyond roughly 1,000 images yields diminishing accuracy returns for this binary dead/live task, so resources are better spent on image quality or scenario diversity.
  • Model selection checklists for aquaculture edge AI should list data volume, hardware class, and required FPS as co-equal criteria with architectural generation.

Reading between the lines

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

  • The same data-efficiency gap may appear in other sparse-object aquaculture tasks (feed pellets, early-stage juveniles) that also rely on one-to-one assignment heads.
  • If ONNX Runtime is replaced by TensorRT or TFLite on the same Pi-class hardware, the relative CPU ranking of NMS-free versus NMS-based heads could shift and should be re-measured.
  • Commercial operators running many tanks may find that the storage and initialization advantages of the smallest YOLO26n ONNX export matter more than raw FPS once multi-camera fleets are considered.
  • The finding that GPU speed orderings reverse on CPU hardware is likely generalizable beyond aquaculture to any edge-vision domain that still ships NMS-based detectors.
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

2 major / 4 minor

Summary. This paper benchmarks the new Ultralytics YOLO26 (NMS-free, edge-oriented) against YOLOv5u, YOLOv8 and YOLO11 at nano/small/medium scales on a binary fish-mortality detection task in a semi-commercial RAS. Twelve models are compared on full-dataset accuracy (mAP50/mAP50-95/precision/recall), data efficiency via learning curves over seven stratified subset sizes, GPU (A100) forward-pass/post-process/end-to-end latency, and ONNX CPU inference on a cooled Raspberry Pi 5. The central empirical claim is that, with the full 2 800-image training set, mAP50 differs by only 1.04 points across all twelve variants, so architectural generation alone does not determine final accuracy; instead data availability, target hardware and required FPS must be considered jointly. YOLOv8 is most data-efficient (≥90 % mAP50 at 400 images), while YOLO26n is fastest on the Pi 5 (7.51 FPS) and YOLOv5mu wins among medium models on CPU.

Significance. If the tabulated rankings hold, the work supplies a concrete, deployment-oriented decision framework for practitioners choosing lightweight detectors for on-farm aquaculture edge devices. Strengths include a fair Ultralytics-only comparison under identical hyper-parameters and fixed splits, dual-platform (A100 + Raspberry Pi 5) timing that exposes GPU-to-CPU rank reversals, systematic learning-curve analysis, and the promised public release of RAS-MortDB plus all twelve .pt/.onnx weights. These elements make the modest claim falsifiable and immediately useful beyond the single-species mortality use-case.

major comments (2)
  1. [Table 3 / §3.1] Table 3 and §3.1.1–3.1.2: the headline claim that architectural generation has “little influence” rests on a 1.04-point mAP50 spread obtained from single training runs (fixed seed 42, one restart per configuration). The authors themselves note that differences “fall well within normal run-to-run training variance,” yet no multi-seed standard deviations or confidence intervals are reported. Without at least three independent full-dataset restarts (or bootstrap estimates), it is impossible to judge whether the observed spread is statistically smaller than training stochasticity; the same caveat applies to the learning-curve thresholds in Figure 5.
  2. [§2.5 / Table 5] §2.5 and Table 5: edge rankings (and the claimed validation of YOLO26’s CPU advantage) are derived from a single actively-cooled Raspberry Pi 5 running ONNX Runtime only. Absolute FPS and the medium-tier reversal (YOLOv5mu > YOLO26m) may change under TFLite, OpenVINO, different ARM SoCs, or thermal throttling. Because the paper’s joint-selection recommendation hinges on hardware-specific behaviour, a second device or runtime would materially strengthen the deployment claim; the current single-platform result remains a useful but narrow data point.
minor comments (4)
  1. [throughout] Inconsistent naming: the text alternates between “YOLO11” and “YOLOv11” (and occasionally “YOLOv5” vs “YOLOv5u”). Standardise on the Ultralytics designations used in Table 1.
  2. [Figure 4] Figure 4 panels lack error bars or shaded ranges; even if only single runs were performed, a note in the caption would help readers interpret the apparent saturation after 1 000 images.
  3. [Introduction / References] Minor typos: “sea food” (p. 2), “intensi ve”, “Fas ter R-CNN”, and a duplicated “Jocher, G.; Qiu, J.” in reference 33. Also, several “accessed on 15 April 2026” dates appear future-dated relative to typical arXiv practice.
  4. [§3.2 / Table 4] Table 4 residual-latency discussion (§3.2) would be clearer if the exact formula Avg − Infer − Post were stated once and the values tabulated.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: pure empirical head-to-head benchmark; measured mAP/FPS/learning-curve numbers do not reduce to inputs by construction.

full rationale

The paper is a deployment-oriented empirical comparison of twelve Ultralytics YOLO variants (YOLOv5u/v8/11/26 imes n/s/m) on a fixed RAS mortality dataset. Detection accuracy (mAP50/mAP50-95), data-efficiency learning curves (seven subset sizes), GPU end-to-end latency, and Raspberry Pi 5 ONNX FPS are obtained by training and timing the models under standardized hyperparameters; none of these quantities is defined in terms of the others or fitted then re-labeled as a prediction. Pretrained COCO weights and Ultralytics defaults are external starting points, not circular. Self-citations (MortCam hardware/protocol [16], prior YOLO aquaculture review [29], image-quality study [10]) supply experimental context and the camera system; they do not define or force the new comparative rankings or the joint-selection conclusion. No uniqueness theorem, ansatz, or renaming of a known result is load-bearing. The derivation chain is therefore self-contained measurement against external hardware and a held-out test set.

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

Empirical computer-vision benchmark; load-bearing content is measured performance under stated experimental conditions rather than free parameters or invented physical entities. The few free choices are standard training hyper-parameters and the operational definition of the ‘dead’ class.

free parameters (3)
  • training epochs / early-stop patience = 100 / 50
    Capped at 100 epochs with patience 50; affects convergence time and final mAP for data-scarce regimes.
  • input resolution and batch size = 640 / 32
    Fixed at 640 px and batch 32 for all models; influences both accuracy and measured latency.
  • learning-curve subset sizes = 100,200,400,700,1000,1400,2800
    Seven discrete sizes (100–2800) chosen by authors to map data efficiency; the 400-image and 1000-image thresholds that drive the main claim are therefore design choices.
assumptions (3)
  • domain assumption Dead fish are visually defined by belly-up floating posture and can be reliably distinguished from live fish by a human annotator under the camera geometry used.
    Ground-truth labels rest on this operational definition (§2.2.2); if posture cues fail for other species or life stages the accuracy numbers do not transfer.
  • domain assumption Ultralytics default automatic optimizer and augmentation pipeline produce fair cross-architecture comparisons when all other settings are locked.
    Stated in §2.3.2; any architecture-specific hyper-parameter sensitivity is absorbed into the measured differences.
  • ad hoc to paper ONNX Runtime on a single actively-cooled Raspberry Pi 5 is a representative proxy for resource-constrained CPU edge deployment in aquaculture.
    Chosen hardware and runtime in §2.5; alternative runtimes (TFLite, OpenVINO) or boards are not tested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Does YOLO26 Truly Offer Advantages Over Its Predecessors for Edge Deployment? A Benchmark Study in Aquaculture." pith.science (2026). https://pith.science/paper/MKMCOOR3

@misc{pith2026260709835,
  author       = {Pith},
  title        = {Pith review of: Does YOLO26 Truly Offer Advantages Over Its Predecessors for Edge Deployment? A Benchmark Study in Aquaculture},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MKMCOOR3}},
  note         = {Machine review of arXiv:2607.09835}
}
read the original abstract

The recently introduced YOLO26 architecture incorporates NMS-free end-to-end inference and is optimized for deployment on resource-constrained CPU-based devices, making it well-suited for edge-based aquaculture applications. However, its performance, operational efficiency, and deployment suitability have not been systematically validated in aquaculture-specific scenarios. This study presents a comprehensive benchmark of YOLO26 against three Ultralytics predecessors (YOLOv5u, YOLOv8, and YOLO11) across nano, small, and medium model scales for fish mortality detection, a critical indicator of fish population health and welfare. Twelve model variants were evaluated for detection accuracy, training efficiency across seven dataset sizes, and inference performance on high-performance NVIDIA A100 GPUs and a CPU-only Raspberry Pi 5 edge platform. All models achieved comparable performance on the full dataset, with mAP50 differing by only 1.04 percentage points, indicating that architectural generation has little influence on final detection accuracy when sufficient training data are available. However, clear trade-offs emerged in data efficiency and deployment performance. YOLOv8 achieved 90% mAP50 with only 400 training images, whereas the YOLO26 nano and small variants required 1,000 images to reach comparable accuracy. Conversely, YOLO26n achieved the highest inference speed on the Raspberry Pi 5 (7.51 FPS), while YOLOv5mu outperformed all contemporary medium-scale architectures on CPU-based hardware. These results show that architectural novelty alone is insufficient for model selection and that training data availability, target hardware, and inference requirements should be considered jointly when selecting object detection models for practical edge AI deployment in aquaculture.

Figures

Figures reproduced from arXiv: 2607.09835 by the authors.

Figure 1
Figure 1. YOLO26 architecture, illustrating the end [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. The (a) MortCam was installed above the bottom drain plate of a semi [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 24 canonical work pages

  1. [1]

    World Aquaculture 2020 – A Brief Overview; FAO Fisheries and Aquaculture Circular No

    Bartley, D.M. World Aquaculture 2020 – A Brief Overview; FAO Fisheries and Aquaculture Circular No. 1233; Food and Agriculture Organization of the United Nations: Rome, Italy, 2022. https://doi.org/10.4060/cb7669en

    work page doi:10.4060/cb7669en 2020

  2. [2]

    Recirculating aquaculture systems (RAS): Environmental solution and climate change adaptation

    Ahmed, N.; Turchini, G.M. Recirculating aquaculture systems (RAS): Environmental solution and climate change adaptation. J. Clean. Prod. 2021, 297, 126604. https://doi.org/10.1016/j.jclepro.2021.126604

    work page doi:10.1016/j.jclepro.2021.126604 2021

  3. [3]

    Environmental performance of marine net-pen aquaculture in the United States

    Rust, M.B.; Amos, K.H.; Bagwill, A.L.; Dickhoff, W.W.; Juarez, L.M.; Price, C.S.; Morris, J.A., Jr.; Rubino, M.C. Environmental performance of marine net-pen aquaculture in the United States. Fisheries 2014, 39, 508–524. https://doi.org/10.1080/03632415.2014.966818

    work page doi:10.1080/03632415.2014.966818 2014

  4. [4]

    Aquaculture: Global status and trends

    Bostock, J.; McAndrew, B.; Richards, R.; Jauncey, K.; Telfer, T.; Lorenzen, K.; Little, D.; Ross, L.; Handisyde, N.; Gatward, I.; Corner, R. Aquaculture: Global status and trends. Philos. Trans. R. Soc. B Biol. Sci. 2010, 365, 2897–2912. https://doi.org/10.1098/rstb.2010.0170

    work page doi:10.1098/rstb.2010.0170 2010

  5. [5]

    Miao, W.; Lal, K.K. (Eds.) Sustainable Intensification of Aquaculture in the Asia-Pacific Region: Documentation of Successful Practices; Food and Agriculture Organization of the United Nations: Bangkok, Thailand, 2016

    2016

  6. [6]

    The economics of recirculating aquaculture systems

    Engle, C.R. The economics of recirculating aquaculture systems. J. World Aquacult. Soc. 2023, 54, 782–785. https://doi.org/10.1111/jwas.13004

    work page doi:10.1111/jwas.13004 2023

  7. [7]

    Precision aquaculture

    O’Donncha, F.; Grant, J. Precision aquaculture. IEEE Internet Things Mag. 2019, 2, 26–30. https://doi.org/10.1109/IOTM.0001.1900033

    work page doi:10.1109/iotm.0001.1900033 2019

  8. [8]

    Precision fish farming: A new framework to improve production in aquaculture

    Føre, M.; Frank, K.; Norton, T.; Svendsen, E.; Alfredsen, J.A.; Dempster, T.; Berckmans, D. Precision fish farming: A new framework to improve production in aquaculture. Biosyst. Eng. 2018, 173, 176–193. https://doi.org/10.1016/j.biosystemseng.2017.10.014

    work page doi:10.1016/j.biosystemseng.2017.10.014 2018

Show all 40 references
  1. [9]

    Detection of residual feed in aquaculture using YOLO and Mask RCNN

    Hu, H.; Tang, C.; Shi, C.; Qian, Y . Detection of residual feed in aquaculture using YOLO and Mask RCNN. Aquac. Eng. 2023, 100, 102304. https://doi.org/10.1016/j.aquaeng.2022.102304

    2023 doi

  2. [10]

    Effects of image data quality on a convolutional neural network trained in-tank fish detection model for recirculating aquaculture systems

    Ranjan, R.; Sharrer, K.; Tsukuda, S.; Good, C. Effects of image data quality on a convolutional neural network trained in-tank fish detection model for recirculating aquaculture systems. Comput. Electron. Agric. 2023, 205, 107644. https://doi.org/10.1016/j.compag.2023.107644

    2023 doi

  3. [11]

    Real-time detection of uneaten feed pellets in underwater images for aquaculture using an improved YOLO-V4 network

    Hu, X.; Liu, Y .; Zhao, Z.; Liu, J.; Yang, X.; Sun, C.; Chen, S.; Li, B.; Zhou, C. Real-time detection of uneaten feed pellets in underwater images for aquaculture using an improved YOLO-V4 network. Comput. Electron. Agric. 2021, 185, 106135. https://doi.org/10.1016/j.compag.2...

    2021 doi

  4. [12]

    Real -time detection and tracking of fish abnormal behavior based on improved YOLOV5 and SiamRPN++

    Wang, H.; Zhang, S.; Zhao, S.; Wang, Q.; Li, D.; Zhao, R. Real -time detection and tracking of fish abnormal behavior based on improved YOLOV5 and SiamRPN++. Comput. Electron. Agric. 2022, 192, 106512. https://doi.org/10.1016/j.compag.2021.106512

    2022 doi

  5. [13]

    Abnormal behavior monitoring method of Larimichthys crocea in recirculating aquaculture system based on computer vision

    Wang, Z.; Zhang, X.; Su, Y .; Li, W.; Yin, X.; Li, Z.; Ying, Y .; Wang, J.; Wu, J.; Miao, F.; Zhao, K. Abnormal behavior monitoring method of Larimichthys crocea in recirculating aquaculture system based on computer vision. Sensors 2023, 23, 2835. https://doi.org/10.3390/s23052835

    2023 doi

  6. [14]

    Fully automatic system for fish biomass estimation based on deep neural network

    Zhang, T.; Yang, Y .; Liu, Y .; Liu, C.; Zhao, R.; Li, D.; Shi, C. Fully automatic system for fish biomass estimation based on deep neural network. Ecol. Inform. 2024, 79, 102399. https://doi.org/10.1016/j.ecoinf.2023.102399

    2024 doi

  7. [15]

    Rapid detection of fish with SVC symptoms based on machine vision combined with a NAM-YOLO v7 hybrid model

    Cai, Y .; Yao, Z.; Jiang, H.; Qin, W.; Xiao, J.; Huang, X.; Pan, J.; Feng, H. Rapid detection of fish with SVC symptoms based on machine vision combined with a NAM-YOLO v7 hybrid model. Aquaculture 2024, 582, 740558. https://doi.org/10.1016/j.aquaculture.2024.740558

    2024 doi

  8. [16]

    MortCam: An artificial intelligence-aided fish mortality detection and alert system for recirculating aquaculture

    Ranjan, R.; Sharrer, K.; Tsukuda, S.; Good, C. MortCam: An artificial intelligence-aided fish mortality detection and alert system for recirculating aquaculture. Aquac. Eng. 2023, 102, 102341. https://doi.org/10.1016/j.aquaeng.2023.102341

    2023 doi

  9. [17]

    An automated lightweight approach for detecting dead fish in a recirculating aquaculture system

    Zhou, C.; Wang, C.; Sun, D.; Hu, J.; Ye, H. An automated lightweight approach for detecting dead fish in a recirculating aquaculture system. Aquaculture 2025, 594, 741433. https://doi.org/10.1016/j.aquaculture.2024.741433

    2025 doi

  10. [18]

    Inspection operations and hole detection in fish net cages through a hybrid underwater intervention system using deep learning techniques

    López- Barajas, S.; Sanz, P.J.; Marín-Prades, R.; Gómez-Espinosa, A.; González-García, J.; Echagüe, J. Inspection operations and hole detection in fish net cages through a hybrid underwater intervention system using deep learning techniques. J. Mar. Sci. Eng. 2024, 12, 80. htt...

    2024 doi

  11. [19]

    Faster R -CNN: Towards real-time object detection with region proposal networks

    Ren, S.; He, K.; Girshick, R.; Sun, J. Faster R -CNN: Towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. 2017, 39, 1137–1149. https://doi.org/10.1109/TPAMI.2016.2577031

    2017 doi

  12. [20]

    Rich feature hierarchies for accurate object detection and semantic segmentation

    Girshick, R.B.; Donahue, J.; Darrell, T.; Malik, J. Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Columbus, OH, USA, 23–28 June 2014; pp. 580–587

    2014

  13. [21]

    You only look once: Unified, real-time object detection

    Redmon, J.; Divvala, S.; Girshick, R.B.; Farhadi, A. You only look once: Unified, real-time object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV , USA, 27–30 June 2016; pp. 779–788

    2016

  14. [22]

    Making Faster R-CNN Faster! Available online: https://jkjung-avt.github.io/making-frcn- faster/ (ac cessed on 15 April 2026 )

    Jung, J.K. Making Faster R-CNN Faster! Available online: https://jkjung-avt.github.io/making-frcn- faster/ (ac cessed on 15 April 2026 )

    2026

  15. [23]

    Progress in object detection: An in- depth analysis of methods and use cases

    Tasnim, S.; Wang, Q. Progress in object detection: An in- depth analysis of methods and use cases. Eur. J. Electr. Eng. Comput. Sci. 2023, 7, 39–45. https://doi.org/10.24018/ejece.2023.7.4.537

    2023 doi

  16. [24]

    YOLOv3: An incremental improvement

    Redmon, J.; Farhadi, A. YOLOv3: An incremental improvement. arXiv 2018, arXiv:1804.02767. https://doi.org/10.48550/arXiv.1804.02767

    2018 doi arXiv

  17. [25]

    Ultralytics YOLOv5 ; Ultralytics, 2020

    Jocher, G. Ultralytics YOLOv5 ; Ultralytics, 2020. Available online: https://docs.ultralytics.com/models/yolov5 (accessed on 15 April 2026)

    2020

  18. [26]

    Chaurasia, A

    Jocher, G.; Qiu, J. ; Chaurasia, A. Ultralytics YOLOv8; Ultralytics, 2023. Available online: https://docs.ultralytics.com/models/yolov8 (accessed on 15 April 2026)

    2023

  19. [27]

    A novel detection model and platform for dead juvenile fish from the perspective of multi-task

    Zhang, P.; Zheng, J.; Gao, L.; Li, P.; Long, H.; Liu, H.; Li, D. A novel detection model and platform for dead juvenile fish from the perspective of multi-task. Multimed. Tools Appl. 2024, 83, 24961– 24981. https://doi.org/10.1007/s11042-023-16370-1

    2024 doi

  20. [28]

    Real -time detection of dead fish for unmanned aquaculture by YOLOv8-based UA V

    Zhang, H.; Tian, Z.; Liu, L.; Liang, H.; Feng, J.; Zeng, L. Real -time detection of dead fish for unmanned aquaculture by YOLOv8-based UA V . Aquaculture 2025, 595, 741551. https://doi.org/10.1016/j.aquaculture.2024.741551

    2025 doi

  21. [29]

    YOLO in precision aquaculture: A decadal bibliometric and systematic review of applications, architectural adaptations, and deployment challenges

    Ranjan, R. YOLO in precision aquaculture: A decadal bibliometric and systematic review of applications, architectural adaptations, and deployment challenges. J. Agric. Food Res. 2026, 28, 102982. https://doi.org/10.1016/j.jafr.2026.102982

    2026 doi

  22. [30]

    Analyzing fish detection and classification in IoT-based aquatic ecosystems through deep learning

    Mohd Rahman, M.I.; Abdul Razak, A.F.; Abdul Majeed, A.P.P.; Musa, R.M.; Abdul Jalil, A.A.; Mohd Khairuddin, I.; Abdullah, M.A.; Mohd Razman, M.A. Analyzing fish detection and classification in IoT-based aquatic ecosystems through deep learning. PeerJ Comput. Sci. 2026, 12, e34...

    2026 doi

  23. [31]

    IoT-enabled communication network for real-time disease alerts in smart aquaculture systems

    Veerappan, S.; Arvinth, N. IoT-enabled communication network for real-time disease alerts in smart aquaculture systems. J. Intell. Assist. Commun. Technol. 2025, 1, 1–7

    2025

  24. [32]

    Ultralytics YOLO26 ; Ultralytics, 2026

    Jocher, G.; Qiu, J. Ultralytics YOLO26 ; Ultralytics, 2026. Available online: https://docs.ultralytics.com/models/yolo26 (accessed on 15 April 2026)

    2026

  25. [33]

    Jocher, G.; Qiu, J

    Jocher, G.; Qiu, J. Jocher, G.; Qiu, J. Ultralytics YOLO11 ; Ultralytics, 2024. Available online: https://docs.ultralytics.com/models/yolo11 (accessed on 15 April 2026)

    2024

  26. [34]

    Ultralytics YOLO evolution: An overview of YOLO26, YOLO11, YOLOv8 and YOLOv5 object detectors for computer vision and pattern recognition

    Sapkota, R.; Karkee, M. Ultralytics YOLO evolution: An overview of YOLO26, YOLO11, YOLOv8 and YOLOv5 object detectors for computer vision and pattern recognition. arXiv 2025, arXiv:2510.09653. https://doi.org/10.48550/arXiv.2510.09653

    2025 doi

  27. [35]

    YOLO26: Key architectural enhancements and performance benchmarking for real-time object detection

    Sapkota, R.; Cheppally, R.H.; Sharda, A.; Karkee, M. YOLO26: Key architectural enhancements and performance benchmarking for real-time object detection. arXiv 2025, arXiv:2509.25164. https://doi.org/10.48550/arXiv.2509.25164

    2025 doi

  28. [36]

    YOLOv11: An overview of the key architectural enhancements

    Khanam, R.; Hussain, M. YOLOv11: An overview of the key architectural enhancements. arXiv 2024, arXiv:2410.17725. https://doi.org/10.48550/arXiv.2410.17725

    2024 doi arXiv

  29. [37]

    YOLOv10: Real-time end-to-end object detection

    Wang, A.; Chen, H.; Liu, L.; Chen, K.; Lin, Z.; Han, J.; Ding, G. YOLOv10: Real-time end-to-end object detection. Adv. Neural Inf. Process. Syst. 2024, 37, 107984–108011. https://doi.org/10.52202/079017-3429

    2024 doi

  30. [38]

    Improving smart home surveillance through YOLO model with transfer learning and quantization for enhanced accuracy and efficiency

    Dalal, S.; Lilhore, U.K.; Sharma, N.; Arora, S.; Simaiya, S.; Ayadi, M.; Almujally, N.A.; Ksibi, A. Improving smart home surveillance through YOLO model with transfer learning and quantization for enhanced accuracy and efficiency. PeerJ Comput. Sci. 2024, 10, e1939. https://do...

    2024 doi

  31. [39]

    YOLO26: An analysis of NMS-free end to end framework for real- time object detection

    Chakrabarty, S. YOLO26: An analysis of NMS-free end to end framework for real- time object detection. arXiv 2026, arXiv:2601.12882. https://doi.org/10.48550/arXiv.2601.12882

    2026 doi

  32. [40]

    Accelerating deep learning model inference on Arm CPUs with ultra-low bit quantization and runtime

    Ashfaq, S.; AskariHemmat, M.; Sah, S.; Saboori, E.; Mastropietro, O.; Hoffman, A. Accelerating deep learning model inference on Arm CPUs with ultra-low bit quantization and runtime. arXiv 2022, arXiv:2207.08820. https://doi.org/10.48550/arXiv.2207.08820

    2022 doi arXiv

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.