--- license: cc-by-4.0 task_categories: - image-classification paperswithcode_id: sykezooscan2024 pretty_name: 'SYKE-plankton_ZooScan_2024' dataset_info: features: - name: image dtype: image - name: label dtype: class_label: names: '0': Bivalvia '1': Bivalvia_multiple '2': Bosmina_sp '3': Bubbles '4': Ceriodaphnia_sp '5': Copepoda_calanoida '6': Copepoda_cyclopoida '7': Copepoda_nauplius '8': Daphnia_sp '9': Eggs '10': Evadne_sp '11': Fibers_etc '12': Fish_eggs '13': Gastropoda '14': Harpacticoida '15': Mysis_sp '16': Podon_sp '17': Polychaeta '18': Sessilia '19': Synchaeta_sp splits: - name: train num_bytes: 62783630 num_examples: 22753 download_size: 68597891 dataset_size: 62783630 description: "The SYKE-plankton_ZooScan_2024 dataset consists of over 24k expert-labeled\ \ single-specimen zooplankton images. The dataset was \nacquired using the ZooScan\ \ instrument applied to water samples collected from the Baltic Sea. The data\ \ is divided into 20 classes \nand all images were annotated by an expert taxonomist.\n\ \nWhile the dataset can be used to train and test plankton recognition models\ \ in general, it was original composed to train and test \nopen-set recognition\ \ (OSR) methods. We provide the splits for OSR used in the original publication.\ \ Furthermore, we provide the \ncorresponding OSR splits for the SYKE-plankton_IFCB_2022\ \ dataset published earlier.\n\nIf you use the SYKE-plankton_ZooScan_2024 dataset\ \ in your research, we kindly ask that you reference the following paper:\n\n\ Kareinen, J., Skyttä, A., Eerola, T., Kraft, K., Lensu, L., Suikkanen, S., Lehtiniemi,\ \ M., & Kälviäinen, H. (2024). Open-Set \nPlankton Recognition. Out of Distribution\ \ Generalization in Computer Vision workshop at ECCV 2024.\n\nFor more details\ \ about the data collection and composition, as well as, the comparison of OSR\ \ methods on both dataset, see the \npaper.\n" dataset_name: 'SYKE-plankton_ZooScan_2024 ' citation: "@article{dataset:sykezooscan2024,\n author = {Kareinen, Joona and Skyttä,\ \ Annaliina}, \n title = {SYKE-plankton_ZooScan_2024}, \n howpublished = {\\\ url{https://doi.org/10.23729/fa115087-2698-4aa5-aedd-11e260b9694d}}, \n month\ \ = {8}, \n year = {2024}, \n note = {LUT University} \n}" homepage: https://etsin.fairdata.fi/dataset/6fa42787-9772-41a5-a6fc-0dde489ed908 configs: - config_name: default data_files: - split: train path: data/train-* dataset_description: "The SYKE-plankton_ZooScan_2024 dataset consists of over 24k\ \ expert-labeled single-specimen zooplankton images. The dataset was \nacquired\ \ using the ZooScan instrument applied to water samples collected from the Baltic\ \ Sea. The data is divided into 20 classes \nand all images were annotated by an\ \ expert taxonomist.\n\nWhile the dataset can be used to train and test plankton\ \ recognition models in general, it was original composed to train and test \nopen-set\ \ recognition (OSR) methods. We provide the splits for OSR used in the original\ \ publication. Furthermore, we provide the \ncorresponding OSR splits for the SYKE-plankton_IFCB_2022\ \ dataset published earlier.\n\nIf you use the SYKE-plankton_ZooScan_2024 dataset\ \ in your research, we kindly ask that you reference the following paper:\n\nKareinen,\ \ J., Skyttä, A., Eerola, T., Kraft, K., Lensu, L., Suikkanen, S., Lehtiniemi, M.,\ \ & Kälviäinen, H. (2024). Open-Set \nPlankton Recognition. Out of Distribution\ \ Generalization in Computer Vision workshop at ECCV 2024.\n\nFor more details about\ \ the data collection and composition, as well as, the comparison of OSR methods\ \ on both dataset, see the \npaper.\n" source_url: https://etsin.fairdata.fi/dataset/6fa42787-9772-41a5-a6fc-0dde489ed908 citation_bibtex: "@article{dataset:sykezooscan2024,\n author = {Kareinen, Joona and\ \ Skyttä, Annaliina}, \n title = {SYKE-plankton_ZooScan_2024}, \n howpublished\ \ = {\\url{https://doi.org/10.23729/fa115087-2698-4aa5-aedd-11e260b9694d}}, \n \ \ month = {8}, \n year = {2024}, \n note = {LUT University} \n}" citation_apa: 'Kareinen, J., & Skyttä, A. (2024). SYKE-plankton_ZooScan_2024 (Version 1). LUT University. https://doi.org/10.23729/fa115087- 26984aa5-aedd-11e260b9694d ' hf_dataset_name: sykezooscan2024 hf_org_name: project-oceania report_markdown: "**Samples per class for split `train`**\n ```────────────────────────\ \ Label histogram for train split ─────────────────────────\n0: Bivalvia \ \ ▇▇ 297.00\n1: Bivalvia_multiple 76.00\n2: Bosmina_sp ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇\ \ 6386.00\n3: Bubbles 63.00\n4: Ceriodaphnia_sp ▇▇▇ 579.00\n5:\ \ Copepoda_calanoida ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 9322.00\n\ 6: Copepoda_cyclopoida ▇ 137.00\n7: Copepoda_nauplius ▇▇ 316.00\n8: Daphnia_sp\ \ ▇▇▇▇▇ 814.00\n9: Eggs 9.00\n10: Evadne_sp ▇\ \ 257.00\n11: Fibers_etc ▇▇▇▇ 773.00\n12: Fish_eggs 10.00\n13:\ \ Gastropoda ▇ 111.00\n14: Harpacticoida 36.00\n15: Mysis_sp \ \ 5.00\n16: Podon_sp ▇▇ 424.00\n17: Polychaeta ▇▇▇▇▇▇▇ 1157.00\n\ 18: Sessilia 7.00\n19: Synchaeta_sp ▇▇▇▇▇▇▇▇▇▇▇ 1974.00\n```\n" dataset_means: '[0.9573715016464085, 0.9573715016464085, 0.9573715016464085]' dataset_stds: '[0.13662989121960673, 0.13662989121960673, 0.13662989121960673]' --- # Dataset *SYKE-plankton_ZooScan_2024* The SYKE-plankton_ZooScan_2024 dataset consists of over 24k expert-labeled single-specimen zooplankton images. The dataset was acquired using the ZooScan instrument applied to water samples collected from the Baltic Sea. The data is divided into 20 classes and all images were annotated by an expert taxonomist. While the dataset can be used to train and test plankton recognition models in general, it was original composed to train and test open-set recognition (OSR) methods. We provide the splits for OSR used in the original publication. Furthermore, we provide the corresponding OSR splits for the SYKE-plankton_IFCB_2022 dataset published earlier. If you use the SYKE-plankton_ZooScan_2024 dataset in your research, we kindly ask that you reference the following paper: Kareinen, J., Skyttä, A., Eerola, T., Kraft, K., Lensu, L., Suikkanen, S., Lehtiniemi, M., & Kälviäinen, H. (2024). Open-Set Plankton Recognition. Out of Distribution Generalization in Computer Vision workshop at ECCV 2024. For more details about the data collection and composition, as well as, the comparison of OSR methods on both dataset, see the paper. - **Original dataset available online at:** . - **Original dataset license:** . ## Details - **train split means (RGB):** [0.9573715016464085, 0.9573715016464085, 0.9573715016464085] - **train split standard deviations (RGB):** [0.13662989121960673, 0.13662989121960673, 0.13662989121960673] **Samples per class for split `train`** ```──────────────────────── Label histogram for train split ───────────────────────── 0: Bivalvia ▇▇ 297.00 1: Bivalvia_multiple 76.00 2: Bosmina_sp ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 6386.00 3: Bubbles 63.00 4: Ceriodaphnia_sp ▇▇▇ 579.00 5: Copepoda_calanoida ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 9322.00 6: Copepoda_cyclopoida ▇ 137.00 7: Copepoda_nauplius ▇▇ 316.00 8: Daphnia_sp ▇▇▇▇▇ 814.00 9: Eggs 9.00 10: Evadne_sp ▇ 257.00 11: Fibers_etc ▇▇▇▇ 773.00 12: Fish_eggs 10.00 13: Gastropoda ▇ 111.00 14: Harpacticoida 36.00 15: Mysis_sp 5.00 16: Podon_sp ▇▇ 424.00 17: Polychaeta ▇▇▇▇▇▇▇ 1157.00 18: Sessilia 7.00 19: Synchaeta_sp ▇▇▇▇▇▇▇▇▇▇▇ 1974.00 ``` ## Reference Kareinen, J., & Skyttä, A. (2024). SYKE-plankton_ZooScan_2024 (Version 1). LUT University. https://doi.org/10.23729/fa115087- 26984aa5-aedd-11e260b9694d ### BibTEX ```bibtex @article{dataset:sykezooscan2024, author = {Kareinen, Joona and Skyttä, Annaliina}, title = {SYKE-plankton_ZooScan_2024}, howpublished = {\url{https://doi.org/10.23729/fa115087-2698-4aa5-aedd-11e260b9694d}}, month = {8}, year = {2024}, note = {LUT University} } ``` ## Usage ```python from datasets import load_dataset dataset = load_dataset("project-oceania/sykezooscan2024") ```