Pith. sign in

REVIEW 3 major objections 5 minor 39 references

Deep learning for Plankton and Coral Classification

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Fusing a diverse set of fine-tuned convolutional neural networks into a score-level ensemble improves plankton and coral classification accuracy over every single model and over published baselines on five benchmark datasets.

desk verdict Useful empirical benchmark on plankton/coral classification, but the 'state-of-the-art' claim is oversold and the SFFS protocol needs clarification before the headline numbers can be trusted. read the letter →

arxiv 1908.05489 v2 pith:RPF3ABMZ submitted 2019-08-15 cs.CV

Alessandra Lumini , Loris Nanni , Gianluca Maguolo This is my paper · ORCID
classification cs.CV
keywords ConvolutionalNeuralNetworkFine-tuningPlanktonClassificationCoralEnsemblelearningTransferUnderwaterimageanalysisSFFSclassifierselection
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 claims that a heterogeneous ensemble of convolutional neural networks (CNNs), combined by score-level fusion, reliably beats both the best single network and published baselines on plankton and coral classification. The authors fine-tune ten well-known CNN architectures from ImageNet weights with deliberate variations—different resizing schemes, one-round versus two-round training, incremental epoch checkpoints, and SELU activations—and test on five public datasets: WHOI, ZooScan, and Kaggle for plankton, EILAT and RSMAS for coral. Using a classifier-selection rule called SFFS, they reduce the full 27-network ensemble to 11 networks and still improve on every dataset, reaching F-measures of 0.958, 0.900, 0.927, 0.990, and 0.995 on the five benchmarks. The practical motivation is that marine experts can manually inspect only a small fraction of collected imagery, so accurate automatic classification is a bottleneck for ecosystem monitoring.

What carries the argument

The mechanism carrying the argument is a score-level ensemble of fine-tuned CNN models selected by Sequential Forward Floating Selection (SFFS), a greedy selection procedure adapted to classifiers: it repeatedly adds the model that improves the ensemble most and then drops the weakest member. The members are standard pretrained architectures—AlexNet, GoogleNet, InceptionV3, VGG-16/19, ResNet50/101, DenseNet, MobileNetV2, and NasNet—fine-tuned with variations that create useful disagreement, including different resizing strategies, one-round versus two-round tuning, incremental epoch snapshots, and SELU instead of ReLU activations. SFFS concentrates that diversity into a compact subset of 11 or 3 models, which is what lets the paper claim both higher accuracy and lower memory than the full fusion.

What would settle it

Re-run the baselines named in Tables 7 and 8 (for example FUS_Hand, Gaussian SVM, MKL, DeepL, and EnsHC) on the exact same dataset splits used for the proposed ensembles, with the same preprocessing. If any baseline reaches or exceeds the SFFS(11) F-measures or accuracies under identical folds, the paper's central claim of state-of-the-art performance fails.

Watch

Extended reading notes

Core claim

On its own terms, the central discovery is that the diversity produced by small training variations is a resource worth harvesting through ensembling. Every architecture is fine-tuned from ImageNet weights under one-round, two-round, incremental, or SELU versions, and the score-level fusion of these models outperforms every single model on all five datasets. DenseNet is the best single model, with F-measures of 0.949, 0.878, 0.912, 0.969, and 0.979 on WHOI, ZooScan, Kaggle, EILAT, and RSMAS, while the selected SFFS(11) ensemble reaches 0.958, 0.900, 0.927, 0.990, and 0.995. The paper also shows that the selection step can shrink the ensemble from 27 to 11 networks, or even to 3, with memory dropping from about 5.5 GB to around 2 GB, and that the same fixed ensemble beats published handcrafted-descriptor and deep-learning results on all five benchmarks.

Load-bearing premise

The load-bearing premise is that the published baseline results were obtained under the same training and test conditions—same folds, splits, and preprocessing—so the ensemble's advantage reflects the method and not an easier evaluation setup.

Editorial extensions

If this is right

  • A fixed recipe of pretrained CNNs, small training variations, and score fusion can reach reported state-of-the-art results on multiple plankton and coral datasets without dataset-specific preprocessing, augmentation, or architecture search.
  • A selected ensemble of 11 models replaces the full 27-model fusion with a much smaller memory footprint while still improving F-measure on all five datasets.
  • Even a single architecture such as DenseNet can form a strong ensemble from its one-round, two-round, and resizing variants, so complete architectural diversity is not strictly required.
  • Training SVMs on CNN-extracted features adds only negligible gains over the CNN's own scores, so the ensemble's benefit comes from combining whole models rather than from changing the final classifier.

Reading between the lines

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

  • The paper's own tables flag that two baselines used different protocols—OPT with a 75/25 split and EnsHC with 10-fold cross-validation—so a strict head-to-head rerun under identical folds is needed before the state-of-the-art margin can be taken at face value.
  • Because the largest gains occur on the harder, class-imbalanced datasets, the same cheap-diversity ensemble recipe is a plausible starting point for other fine-grained marine vision tasks, such as fish species or benthic invertebrate classification.
  • The reported memory savings count stored weights; running all 11 networks still requires activations and forward passes, so deployment on small underwater vehicles would likely need additional compression or distillation.
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 / 5 minor

Summary. The paper studies ensembles of ImageNet-pretrained CNN architectures, fine-tuned on five underwater image datasets (WHOI, ZooScan, Kaggle, EILAT, RSMAS), for plankton and coral classification. It compares single architectures, ensemble rules (Fus_*, DN_*), transfer learning to SVMs, and classifier-selection strategies SFFS and WS. The headline result is that a heterogeneous ensemble reduced by SFFS to 11 CNNs achieves F-measures of 0.958, 0.900, 0.927, 0.990, and 0.995 on the five datasets, outperforming the best single model (DenseNet) and previously published state-of-the-art methods. The paper also reports memory costs for the reduced ensembles and claims that the same pipeline works 'out-of-the-box' across datasets without per-dataset tuning.

Significance. If the reported results are valid estimates of generalization, the paper is a useful empirical contribution: it provides a broad comparison of fine-tuned CNN architectures and shows that ensembles of diverse architectures consistently improve over single models in three plankton and two coral classification tasks. The release of MATLAB source code for reproduction is a positive feature. The main value is practical, giving practitioners a concrete recipe (diverse architectures, incremental training, SFFS selection) for improving accuracy in similar small-to-medium marine image classification problems. However, the significance of the central 'substantial improvement' claim depends critically on the validity of the SFFS evaluation protocol and on the comparability of the state-of-the-art baselines, both of which are open concerns in the current manuscript.

major comments (3)
  1. [§3, Table 6 and final paragraph of Section 3] The headline result SFFS(11) appears to be produced by the sentence 'Finally, we ran SFFS using all the five datasets. In this way we find an ensemble of 11 CNNs...' This is in direct tension with the leave-one-out-dataset protocol described for SFFS in Section 2 (Fig. 2). If the same five datasets were used both for selecting the 11-member ensemble and for evaluating the F-measures in Table 6, then the Table 6 numbers are selection results on training data, not unbiased predictions on unseen test data. Because SFFS(11) is the basis of the paper's 'substantial performance improvement' conclusion, the authors must clarify whether the 11 models were selected on held-out data, or add a truly held-out evaluation of the final ensemble (e.g., nested cross-validation or a separate split).
  2. [§3, Tables 7 and 8]
  3. [Tables 1–6] All results are reported as point estimates with no error bars, standard deviations, or repeated-run information. Since the key claimed improvements (e.g., SFFS(11) versus Fus_2R+Fus_1R, or versus the best single model) are often on the order of 0.001–0.005, the reader cannot determine whether these differences are statistically meaningful. Reporting means and standard deviations across the folds (or across multiple training runs) is necessary to support the 'substantial improvement' claim.
minor comments (5)
  1. [§2, Fig. 2 pseudo-code] In the backtracking step of the SFFS pseudo-code, the line 'RS:=RS+smax' appears to be a typo; after removing smin from SS, the algorithm should add smin back to RS, i.e., 'RS:=RS+smin'.
  2. [§2–§3, model naming] The model list in Section 2 refers to 'DenseNet', but the final SFFS list includes 'densenet201_1R_Pad/Tile' and 'densenet201_2R'. Please clarify whether DenseNet in all experiments is DenseNet-201, and keep the naming consistent across tables and the final ensemble list.
  3. [§3, Fig. 4] The y-axis of Fig. 4 starts at 0.86, which visually exaggerates the differences between methods; consider starting at 0.80 or showing change relative to a baseline.
  4. [§3, performance indicators] The formula for multi-class F-measure and accuracy is written with a per-class index c, but the averaging notation is ambiguous. Please define P_C and R_C explicitly and state whether the reported F-measure is macro-averaged (as it appears to be).
  5. [§3, Dataset split description] The text says 'the class distribution has not been maintained when splitting the dataset' but then describes standard protocols (e.g., '2-fold cross validation', '5-fold cross validation'). Please state whether the folds were stratified or not, since this affects reproducibility.

Circularity Check

1 steps flagged · score 6.0 of 10

SFFS(11) is selected using all five datasets and then reported on the same datasets, making its Table 6 F-measures in-sample selection scores rather than held-out predictions.

  1. fitted input called prediction [Section 2 (SFFS description) and Section 3 (final paragraph, Table 6)]
    "Since SFFS requires a training phase, in order to select the best suited models, we perform a leave-one-out-dataset selection. Finally, we ran SFFS using all the five datasets. In this way we find an ensemble of 11 CNNs that includes: 1. AlexNet_1R_Pad/Tile 2. AlexNet_2R 3. vgg16_1R_SqR 4. vgg16_INC 5. googlenet_1R_Pad/Tile 6. googlenet_INC 7. resnet50_2R 8. inceptionv3_1R_Pad/Tile 9. densenet201_1R_Pad/Tile 10. densenet201_2R 11. nasnetlarge_1R_SqR. Table 6: SFFS(11 classifiers) 0.958 0.900 0.927 0.990 0.995."

    SFFS is defined as a trained selection rule that picks ensemble members by their incremental performance. The paper's own Fig. 2 protocol avoids circularity by selecting on a leave-one-out-dataset basis, excluding the tested dataset. But the final paragraph replaces that protocol with 'we ran SFFS using all the five datasets' to obtain the 11-CNN ensemble, and Table 6 reports that ensemble's F-measures on the same five datasets. The ensemble composition is therefore fitted to the very evaluation partitions whose scores are presented as SFFS(11). Those numbers are in-sample selection scores, not predictions on unseen data, so the claimed improvement of SFFS(11) over single models and fixed ensembles is partly forced by construction.

full rationale

This paper is primarily an empirical benchmark study rather than a derivation, so most circularity patterns do not apply. The single-model results and the fixed ensemble results (Fus_1R, Fus_2R, DN_1R+2R, etc.) are self-contained: models are trained on training splits and evaluated on test splits, and no fitted normalization or definitional identity forces those numbers. The one load-bearing circular step is the final SFFS(11) result. The paper defines SFFS as a trained rule that selects classifiers by incremental performance, gives a leave-one-out-dataset protocol in Fig. 2, but then states that the final 11-CNN ensemble was obtained by running SFFS on all five datasets. Table 6 then reports the F-measures of that selected ensemble on the same five datasets. By construction, those numbers maximize in-sample selection performance and are not unbiased held-out estimates of generalization. The central claim of substantial improvement over state-of-the-art is partially supported by the independently composed Fus_2R+Fus_1R ensemble, so the paper retains independent content, but its headline 'best' result is compromised. Self-citations in the text, including the authors' earlier plankton paper and the FUS_Hand baseline, are not load-bearing premises of the ensemble construction. The protocol mismatches noted in Tables 7 and 8 (75/25 split and 10-fold CV for baselines) are comparison-validity concerns rather than circularity and do not affect this score.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The paper adds no new mathematical derivation and no invented entities. Its empirical claims rest on hand-chosen hyperparameters and on standard transfer-learning assumptions. The leave-one-out-dataset protocol for SFFS/WS is a legitimate attempt to avoid test-set selection, but the final all-data SFFS list is not evaluated in a held-out way.

free parameters (5)
  • WS regularization exponent gamma = not reported
    Controls how many classifiers receive nonzero weight in the WS ensemble. The paper says the balance depends on gamma but never states its value, so the exact weighting is not reproducible.
  • Learning rate = 0.001
    Fixed for all fine-tuning runs; chosen by hand rather than optimized per dataset.
  • Training epochs = 30 for 1R/2R; 45 for INC
    Number of fine-tuning epochs is set by hand, and incremental training stores a network every 3 epochs.
  • Mini-batch size = 16 to 128 depending on model
    Assigned per architecture based on memory constraints; affects the optimization trajectory and final accuracy.
  • Ensemble size for SFFS/WS = 11 and 3 classifiers
    The stopping criteria are not derived; the paper reports results for both sizes and uses 11 as the headline result.
assumptions (4)
  • domain assumption ImageNet pre-trained weights transfer to underwater plankton and coral images
    All fine-tuning starts from ImageNet models; if this transfer fails, the central recipe collapses. See Section 2.
  • domain assumption Baseline results from cited papers are comparable because the same testing protocols were used
    The state-of-the-art comparison in Tables 7 and 8 relies on this; the paper asserts it directly, though the table footnotes show protocol differences for two baselines.
  • standard math Standard SGD training with cross-entropy loss and fixed learning rate 0.001 behaves as reported
    Invoked throughout Section 2; no derivation is given, which is normal for empirical deep learning.
  • domain assumption Not maintaining class distribution in splits produces a meaningful dataset-drift evaluation rather than an unfair test
    Section 3 states this choice; it deviates from common stratified cross-validation and affects comparability with baselines that may use stratified splits.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep learning for Plankton and Coral Classification." pith.science (2026). https://pith.science/paper/RPF3ABMZ

@misc{pith2026190805489,
  author       = {Pith},
  title        = {Pith review of: Deep learning for Plankton and Coral Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RPF3ABMZ}},
  note         = {Machine review of arXiv:1908.05489}
}
read the original abstract

Oceans are the essential lifeblood of the Earth: they provide over 70% of the oxygen and over 97% of the water. Plankton and corals are two of the most fundamental components of ocean ecosystems, the former due to their function at many levels of the oceans food chain, the latter because they provide spawning and nursery grounds to many fish populations. Studying and monitoring plankton distribution and coral reefs is vital for environment protection. In the last years there has been a massive proliferation of digital imagery for the monitoring of underwater ecosystems and much research is concentrated on the automated recognition of plankton and corals. In this paper, we present a study about an automated system for monitoring of underwater ecosystems. The system here proposed is based on the fusion of different deep learning methods. We study how to create an ensemble based of different CNN models, fine tuned on several datasets with the aim of exploiting their diversity. The aim of our study is to experiment the possibility of fine-tuning pretrained CNN for underwater imagery analysis, the opportunity of using different datasets for pretraining models, the possibility to design an ensemble using the same architecture with small variations in the training procedure. The experimental results are very encouraging, our experiments performed on 5 well-knowns datasets (3 plankton and 2 coral datasets) show that the fusion of such different CNN models in a heterogeneous ensemble grants a substantial performance improvement with respect to other state-of-the-art approaches in all the tested problems. One of the main contributions of this work is a wide experimental evaluation of famous CNN architectures to report performance of both single CNN and ensemble of CNNs in different problems. Moreover, we show how to create an ensemble which improves the performance of the best single model.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 27 canonical work pages

  1. [1]

    Larkum, R.J

    A.W.D. Larkum, R.J. Orth, C.M. Duarte, Seagrasses: Biology, ecology and conservation, 2006. doi:10.1007/978- 1-4020-2983-7

    doi:10.1007/978- 2006

  2. [2]

    Olson, H.M

    R.J. Olson, H.M. Sosik, A submersible imaging-in-flow instrument to analyze nano-and microplankton: Imaging FlowCytobot, Limnol. Oceanogr. Methods. 5 (2007) 195–203. doi:10.4319/lom.2007.5.195

    work page doi:10.4319/lom.2007.5.195 2007

  3. [3]

    Moniruzzaman, S.M.S

    M. Moniruzzaman, S.M.S. Islam, M. Bennamoun, P. Lavery, Deep learning on underwater marine object detection: A survey, in: Lect. Notes Comput. Sci. (Including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), 2017: pp. 150–160. doi:10.1007/978-3-319-70353-4_13

    work page doi:10.1007/978-3-319-70353-4_13 2017

  4. [4]

    H. Qin, X. Li, Z. Yang, M. Shang, When underwater imagery analysis meets deep learning: A solution at the age of big visual data, in: Ocean. 2015 - MTS/IEEE Washingt., 2016

    work page 2015

  5. [5]

    Rathi, S

    D. Rathi, S. Jain, S. Indu, Underwater Fish Species Classification using Convolutional Neural Network and Deep Learning, in: 2017 9th Int. Conf. Adv. Pat tern Recognition, ICAPR 2017, 2018. doi:10.1109/ICAPR.2017.8593044

    work page arXiv 2017

  6. [6]

    F. Zhao, F. Lin, H.S. Seah, Binary SIPPER plankton image classification using random subspace, Neurocomputing. 73 (2010) 1853–1860. doi:10.1016/j.neucom.2009.12.033

    work page doi:10.1016/j.neucom.2009.12.033 2010

  7. [7]

    Zheng, R

    H. Zheng, R. W ang, Z. Yu, N. Wang, Z. Gu, B. Zheng, Automatic plankton image classification combining multiple view features via multiple kernel learning, BMC Bioinformatics. 18 (2017) 1 –18. doi:10.1186/s12859- 017-1954-8

    doi:10.1186/s12859- 2017

  8. [8]

    H. Lee, M. Park, J. Kim, Plankton classifica tion on imbalanced large scale database via convolutional neural networks with transfer learning, in: Proc. - Int. Conf. Image Process. ICIP, 2016: pp. 3713 –3717. doi:10.1109/ICIP.2016.7533053

    arXiv 2016

Show all 39 references
  1. [9]

    J. Dai, R. Wang, H. Zheng, G. Ji, X. Qiao, ZooplanktoNet: Deep convolutional network for zooplankton classification, in: Ocean. 2016 - Shanghai, 2016. doi:10.1109/OCEANSAP.2016.7485680

    2016

  2. [10]

    J. Gu, Z. Wang, J. Kuen, L. Ma, A. Shahroudy, B. Shuai, et al., Recent advances in convolutional neural networks, Pattern Recognit. 77 (2018) 354–377. doi:10.1016/J.PATCOG.2017.10.013

    2018 doi

  3. [11]

    Dieleman, J

    S. Dieleman, J. De Fauw, K. Kavukcuoglu, Exploiting Cyclic Symmetry in Convolutional Neural Networks, CoRR. abs/1602.0 (2016). http://arxiv.org/abs/1602.02660

    2016 arXiv

  4. [12]

    O. Py, H. Hong, S. Zhong zhi, Plankton classification with deep convolutional neural networks, in: 2016 IEEE Inf. Technol. Networking, Electron. Autom. Control Conf., 2016: pp. 132 –136. doi:10.1109/ITNEC.2016.7560334

    2016

  5. [13]

    J. Dai, Z. Yu, H. Zheng, B. Zheng, N. Wang, A Hybrid Convolutional Neural Network for Plankton Classification, in: C.-S. Chen, J. Lu, K. -K. Ma (Eds.), Comput. Vis. -- ACCV 2016 Work., Springer International Publishing, Cham, 2017: pp. 102–114

    2016

  6. [14]

    Bochinski, G

    E. Bochinski, G. Bacha, V. Eiselein, T.J.W. Walles, J.C. Nejstgaard, T. Sikora, Deep Active Learning for In Situ Plankton Classification, in: Z. Zhang, D. Suter, Y. Tian, A. Branzan Albu, N. Sidère, H. Jair Escalante (Eds.), Pattern Recognit. Inf. Forensics, Springer Internati...

    2019

  7. [15]

    Rodrigues, N.S.T

    F.C.M. Rodrigues, N.S.T. Hirata, A.A. Abello, L.T.D. La Cruz, R.M. Lopes, R.H. Jr., Evaluation of Transfer Learning Scenarios in Plankton Image Classification, in: Proc. 13th Int. Jt. Conf. Comput. Vision, Imaging 14 Comput. Graph. Theory Appl. - Vol. 5 VISAP P, SciTePress, 20...

    2018 doi

  8. [16]

    Lumini, L

    A. Lumini, L. Nanni, Deep learning and transfer learning features for plankton classification, Ecol. Inform. (2019). doi:10.1016/j.ecoinf.2019.02.007

    2019 doi

  9. [17]

    Cheng, X

    K. Cheng, X. Cheng, Y. Wang, H. Bi, M.C. Benfield, Enhanced convolutional neural network for plankton identification and enumeration, PLoS One. (2019). doi:10.1371/journal.pone.0219570

    2019 doi

  10. [18]

    A.B.M. N, D. Dharma, Coral reef image/video classification employing novel octa -angled pattern for triangular sub region and pulse coupled convolutional neural network (PCCNN), Multimed. Tools Appl. 77 (2018) 31545– 31579. doi:10.1007/s11042-018-6148-5

    2018 doi

  11. [19]

    Mahmood, M

    A. Mahmood, M. Bennamoun, S. An, F. Sohel, F. Boussaid, R. Hovey, et al., Coral classification with hybrid feature representations, in: Proc. - Int. Conf. Image Process. ICIP, 2016. doi:10.1109/ICIP.2016.7532411

    2016

  12. [20]

    Mahmood, M

    A. Mahmood, M. Bennamoun, S. An, F.A. Sohel, F. Boussaid, R. Hovey, et al., Deep Image Representations for Coral Image Classification, IEEE J. Ocean. Eng. (2019). doi:10.1109/JOE.2017.2786878

    2019

  13. [21]

    Beijbom, T

    O. Beijbom, T. Treibitz, D.I. Kline, G. Eyal, A. Khen, B. Neal, et al., Improving Automated Annotation of Benthic Survey Images Using Wide-band Fluorescence, Sci. Rep. (2016). doi:10.1038/srep23166

    2016 doi

  14. [22]

    Gómez-Ríos, S

    A. Gómez-Ríos, S. Tabik, J. Luengo, A.S.M. Shihavuddin, B. Krawczyk, F. Herrera, Towards highly accurate coral texture images classification using deep convolutional neural networks and data augmentation, Expert Syst. Appl. (2019). doi:10.1016/j.eswa.2018.10.010

    2019 doi

  15. [23]

    Goodfellow, Ian, Bengio, Yoshua, Courville, Deep Learning, MIT Press

    A. Goodfellow, Ian, Bengio, Yoshua, Courville, Deep Learning, MIT Press. (2016). http://www.deeplearningbook.org/

    2016

  16. [24]

    Klambauer, T

    G. Klambauer, T. Unterthiner, A. Mayr, S. Hochreiter, Self -Normalizing Neural Networks, in: NIPS, 2017. http://arxiv.org/abs/1706.02515 (accessed June 19, 2019)

    2017 arXiv

  17. [25]

    Krizhevsky, I

    A. Krizhevsky, I. Sutskever, G.E. Hinton, ImageNet Classification with Deep Convolutional Neural Networks, Adv. Neural Inf. Process. Syst. (2012) 1–9. doi:http://dx.doi.org/10.1016/j.protcy.2014.09.007

    2012 doi

  18. [26]

    Szegedy, W

    C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, et al., Going deeper with convolutions, in: Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., 2015: pp. 1–9. doi:10.1109/CVPR.2015.7298594

    2015

  19. [27]

    Szegedy, V

    C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, Z. Wojna, Rethinking the Inception Architecture for Computer Vision, in: 2016 IEEE Conf. Comput. Vis. Pattern Recognit., 2016: pp. 2818–2826. doi:10.1109/CVPR.2016.308

    2016 doi

  20. [28]

    Simonyan, A

    K. Simonyan, A. Zisserman, Very Deep Convolutional Networks for Large-Scale Image Recognition, Int. Conf. Learn. Represent. (2015) 1–14. doi:10.1016/j.infsof.2008.09.005

    2015 doi

  21. [29]

    K. He, X. Zhang, S. Ren, J. Sun, Deep Residual Learning for Image Recognition, in: 2016 IEEE Conf. Comput. Vis. Pattern Recognit., 2016: pp. 770–778. doi:10.1109/CVPR.2016.90

    2016 doi

  22. [30]

    Huang, Z

    G. Huang, Z. Liu, L. Van Der Maaten, K.Q. Weinberger, Densely connected convolutional networks, in: Proc. - 30th IEEE Conf. Comput. Vis. Pattern Recognition, CVPR 2017, 2017. doi:10.1109/CVPR.2017.243

    2017 doi

  23. [31]

    Sandle r, A

    M. Sandle r, A. Howard, M. Zhu, A. Zhmoginov, L.C. Chen, MobileNetV2: Inverted Residuals and Linear Bottlenecks, in: Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., 2018. doi:10.1109/CVPR.2018.00474

    2018

  24. [32]

    B. Zoph, V. Vasudevan, J. Shlens, Q. V. Le, Learning Transferable Architectures for Scalable Image Recognition, in: Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., 2018. doi:10.1109/CVPR.2018.00907

    2018

  25. [33]

    Pudil, J

    P. Pudil, J. Novovičová, J. Kittl er, Floating search methods in feature selection, Pattern Recognit. Lett. (1994). doi:10.1016/0167-8655(94)90127-9

    1994 doi

  26. [34]

    Sosik, R.J

    H.M. Sosik, R.J. Olson, Automated taxonomic classification of phytoplankton sampled with imaging -in-flow cytometry, Limnol. Oceanogr. Methods. 5 (2007) 204–216. doi:10.4319/lom.2007.5.204

    2007 doi

  27. [35]

    Gorsky, M.D

    G. Gorsky, M.D. Ohman, M. Picheral, S. Gasparini, L. Stemmann, J.B. Romagnan, et al., Digital zooplankton image analysis using the ZooScan integrated system, J. Plankton Res. 32 (2010) 285 –303. doi:10.1093/plankt/fbp124

    2010 doi

  28. [36]

    Gonzalez, E

    P. Gonzalez, E. Alvarez, J. Diez, A. Lopez -Urrutia, J.J. del Coz, Validation methods for plankton image classification systems, Limnol. Oceanogr. Methods. 15 (2017) 221–237. doi:10.1002/lom3.10151

    2017 doi

  29. [37]

    Nanni, S

    L. Nanni, S. Ghidoni, S. Brahnam, Handcrafted vs. non-handcrafted features for computer vision classification, Pattern Recognit. 71 (2017) 158–172. doi:10.1016/j.patcog.2017.05.025

    2017 doi

  30. [38]

    Nanni, A

    L. Nanni, A. Lumini, Ocean Ecosystems Plankton Classification, in: M. Hassaballah, K.M. Hosny (Eds.), Recent Adv. Comput. Vis. Theor. Appl., Springer, 2018

    2018

  31. [39]

    Shihavuddin, N

    A.S.M. Shihavuddin, N. Gracias, R. Garcia, A.C.R. Gleason, B. Gintert, Image-based coral reef classification and thematic mapping, Remote Sens. (2013). doi:10.3390/rs5041809. 15

    2013 doi

Pith tools

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