{
  "fetched_at": "2026-10-24T15:16:15Z",
  "source_url": "https://www.kaggle.com/competitions/fungiclef2022/data",
  "pages": [
    {
      "name": "data-description",
      "content": "The challenge dataset is based on the data from the [Danish Fungi 2020](https://openaccess.thecvf.com/content/WACV2022/papers/Picek_Danish_Fungi_2020_-_Not_Just_Another_Image_Recognition_Dataset_WACV_2022_paper.pdf) dataset, which contains 295,938 training images belonging to 1,604 species observed mostly in Denmark. All training samples passed an expert validation process, guaranteeing high-quality labels. Rich observation metadata about habitat, substrate, time, location, EXIF etc. are provided.\n\nThe test set contains 59,420 observations with 118,676 images and 3,134 species, covering the whole year and includes observations collected across all substrate and habitat types.\n\n*The FungiCLEF2022 competition is based on data submitted to the [Atlas of Danish Fungi.](https://svampe.databasen.org/)*\n\n**Images**\n\n- Development set - [Full Size](http://ptak.felk.cvut.cz/plants/DanishFungiDataset/DF20-train_val.tar.gz) [~110Gb]\n- Development set - Max side Size - 300px (Available bellow) [~6.5Gb]\n\n- Test Images - [Full Size](http://ptak.felk.cvut.cz/plants/DanishFungiDataset/DF21-images.tar.gz) [~50Gb]\n- Test Images (max side size 300px) - Available bellow in folder [DF21-images-300]\n\n**Metadata**\n\n- Train metadata (Available bellow)\n- Test metadata (Available bellow) [DF21-images-300]",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "## Motivation\nAutomatic recognition of fungi species assists mycologists, citizen scientists and nature enthusiasts in species identification in the wild. Its availability supports the collection of valuable biodiversity data. In practice, species identification typically does not depend solely on the visual observation of the specimen but also on other information available to the observer — such as habitat, substrate, location and time. Thanks to rich metadata, precise annotations, and baselines available to all competitors, the challenge provides a benchmark for image recognition with the use of additional information. Moreover, the toxicity of a mushroom can be crucial for the decision of a mushroom picker. We will explore the decision process within the competition beyond the commonly assumed 0/1 cost function.\n\n## Task Description\nGiven the set of fungi observations — multiple photographs of the same individual — and corresponding observation metadata, the goal of the task is to return for each image one species based on the likelihood that they might have been observed at that location.\n\n**Note:** *Since the test set contains multiple out-of the-scope classes. Solution have to handle such classes.*\n\n## Context\n- This competition is held jointly as part of:\n- the LifeCLEF 2022 lab of the CLEF 2022 conference, and of\n- the FGVC9 workshop, organized in conjunction with CVPR 2022 conference.\n\nThe participants are required to participate to the LifeCLEF lab by registering to it using this [form](https://clef2022-labs-registration.dei.unipd.it/) (and checking \"Task 5 - FungiCLEF\" of LifeCLEF). Only registered participants can submit a working-note paper to peer-reviewed LifeCLEF proceedings (CEUR-WS) after the competition ends.\"\n\n*This paper should provide sufficient information to reproduce the final submitted runs.\nOnly participants who submitted a working-note paper will be part of the officially published ranking used for scientific communication.*\n\n## Best Paper Awards\nWe will offer a combined credit grant of 5k USD as part of Microsoft's [AI for Earth](https://www.microsoft.com/en-us/ai/ai-for-earth) program to the best papers (up to 3).\n\nIn order to claim the Azure price, participants are required to register to LifeCLEF and submit a working-note paper.\nThe prices will be given to the teams among the highest ranking on the private leaderboard who will have matched these requirements.\n\nSee the [Rules](https://www.kaggle.com/c/fungiclef2022/rules) section for more details.\n\n## Organizers\n\n**Machine Learning**\n[Lukas Picek](https://www.sites.google.com/view/picekl), Dept. of Cybernetics, FAV, University of West Bohemia, Czechia, lukaspicek@gmail.com\n[Milan Sulc](https://sulc.github.io/), Rossum.ai\n[Jiri Matas](https://cmp.felk.cvut.cz/~matas/), The Center for Machine Perception Dept. of Cybernetics, FEE, Czech Technical University, Czechia\n\n**Mycology**\n[Jacob Heilmann-Clausen](https://globe.ku.dk/staff-list/?pure=en/persons/14882), Center for Macroecology, Evolution and Climate University of Copenhagen, Denmark, jheilmann-clausen@sund.ku.dk\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "## Metric \nThe evaluation metric for this competition is [Mean F1-Score](https://en.wikipedia.org/wiki/F-score). The F1 score, commonly used in information retrieval, measures classification performance using the statistics precision (P) and recall (R).\n\nThe F1 score is not biased by class frequencies and is more suitable for the long-tailed class distributions observed in nature.\n\nPrecision is the ratio of true positives (TP) to all predicted positives (TP + FP). Recall is the ratio of true positives (TP) to all actual positives (TP + FN).\n\nThe F1 metric weights recall and precision equally, and a good retrieval algorithm will maximize both precision and recall simultaneously. Thus, moderately good performance on both will be favoured over extremely good performance on one and poor performance on the other.\n\n## Submission Format\n\nFor each image in the test set, you must predict 1 class label. The CSV file have to contain a header and have the following format:\n```\nObservationId,ClassId\n3305985310,307\n3305985311,597\n3305985312,1060\n3305985315,860\n3305985316,200\n```\n\nThe *ObservationId* column corresponds to the observation ID. \nThe *ClassId* column corresponds to one category id.\n\n**For out-of-the-scope classes use -1 as ClassId.**\n\nYou should have one row for each test observation.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "# Rules\n## General rules\n### One account per participant\nYou cannot sign up to Kaggle from multiple accounts and therefore you cannot submit from multiple accounts.\n### No private sharing outside teams\nPrivately sharing code or data outside of teams is not permitted.\nIt is okay to share code if made available to all participants on the forums.\n### Team mergers\nTeam mergers are allowed and can be performed by the team leader.\nIn order to merge, the combined team must have a total submission count less than or equal to the maximum allowed as of the merge date.\nThe maximum allowed is the number of submissions per day multiplied by the number of days the competition has been running.\n### Team limits\nThere is maximum team size set to 5.\n### Submission limits\nYou may submit a maximum of 2 entries per day.\nYou may select up to 10 final submissions for judging.\n### Competition timeline\nCompetition timeline dates are reflected on the competition’s **Overview > Timeline page**.\n### Competition data access and usage\nThe data used for this competition is aggregated from several sources.\nFor the details of the data licenses, see this page.\nIn particular, the dataset can be used for academic research and education.\n### Use of external data and pre-trained models\nUsing external observation data (such as data from GBIF, etc.) is forbidden.\nOther external data (such as image data) used for model pre-training is authorized as long as it is publicly accessible.\nIn particular, using pre-trained models from standard sources, such as Pytorch hub, torchvision, etc., is authorized.\nIn case of doubt, please contact us!\nIf such data or pre-trained is used, they must be explicitly referenced in the submitted working note papers (see below).\n\n## Official ranking of LifeCLEF - FungiCLEF 2022\nIn order to be part of the official ranking of FungiCLEF 2022 lab which will be published and used in scientific communication, the teams are required to submit a working note paper to LifeCLEF workshop describing the methods implemented which will be peer reviewed.\nThis working note must provide sufficient details to allow to reproduce all final submissions.\nAny run that could not be reproduced thanks to its description in the working notes might be removed from the official publication of the results.\nWorking notes are published within CEUR-WS proceedings, resulting in an assignment of an individual DOI (URN) and an indexing by many bibliography systems including DBLP.\nAccording to the CEUR-WS policies, a single-blind, two stage review of the working notes will be conducted by LifeCLEF organizing committee and external reviewers to ensure quality. As an illustration, LifeCLEF 2021 working notes (task overviews and participant working notes) can be found within [CLEF 2021 CEUR-WS proceedings] (http://ceur-ws.org/Vol-2936/).\n\nBelow the link gives more information about the templates (latex, doc) to use, the mandatory information to report (about the names, affiliations, etc):\n\n- [CLEF instructions for authors](https://clef2022.clef-initiative.eu/index.php?page=Pages/instructions_for_authors.html)\n- [The direct link to submit your working note through easychair](https://easychair.org/my/conference?conf=clef2022)\n\n## Best Paper Awards\nWe will offer a combined credit grant of 5k USD as part of Microsoft's [AI for Earth](https://www.microsoft.com/en-us/ai/ai-for-earth) program to the best papers (up to 3).",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Timeline",
      "content": "## Timeline\nJan 2022: Registration opens for all LifeCLEF challenges\nFeb 2022: Competition start date\n**15 May 2022: Competition Deadline**\n**29 May 2022: Paper submission Deadline — working note papers by participants**\n14 June 2022: Reviews - First Round   |   Notification of Acceptance\n**19 June: CVPR 2022 - FGVC Workshop**\n23 June 2022: Reviews - Second Round  |   Notification of Acceptance \n1 July 2022: camera ready working note papers by participants and organizers\n5-9 Sept 2022: [CLEF 2022](https://clef2022.clef-initiative.eu/) Università di Bologna\n\n*All deadlines are at 11:59 PM UTC on the corresponding day unless otherwise noted.\nThe competition organizers reserve the right to update the contest timeline if they deem it necessary.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Other Resources",
      "content": "### Other Resources\n\n- [Paper](https://openaccess.thecvf.com/content/WACV2022/html/Picek_Danish_Fungi_2020_-_Not_Just_Another_Image_Recognition_Dataset_WACV_2022_paper.html) describing the dataset\n- [LifeCLEF](https://www.imageclef.org/LifeCLEF2022) page for more information about LifeCLEF challenges and working notes submission procedure\n- [ FGVC](https://sites.google.com/view/fgvc9) page for more information about the FGVC9 workshop",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "CLEF2022 – LifeCLEF Workshop",
      "content": "### LifeCLEF Workshop\nLifeCLEF 2022 is an evaluation campaign that is being organized as part of the [CLEF initiative labs](https://clef2022.clef-initiative.eu/). The campaign offers several research tasks that welcome participation from teams around the world.\nCLEF 2022 will be hosted by Università di Bologna, Italy, 5-8 September 2022\nMore details can be found on the CLEF 2022 [website](https://clef2022.clef-initiative.eu/).\n\nSubmitting a working note with the full description of the methods used in each run is mandatory.\nAny run that could not be reproduced thanks to its description in the working notes might be removed from the official publication of the results.\nWorking notes are published within CEUR-WS proceedings, resulting in an assignment of an individual DOI (URN) and an indexing by many bibliography systems including Scopus and DBLP.\n\nAccording to the CEUR-WS policies, a single-blind, two stage review of the working notes will be conducted by LifeCLEF organizing committee and external reviewers to ensure quality. As an illustration, LifeCLEF 2021 working notes (task overviews and participant working notes) can be found within [CLEF 2021 CEUR-WS proceedings] (http://ceur-ws.org/Vol-2936/).\n\n*Selected contributions among the participants, will be invited for publication in the following year in the Springer Lecture Notes in Computer Science (LNCS) together with the annual lab overviews.*",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "CVPR2022 – FGVC Workshop",
      "content": "### FGVC Workshop\nThis competition is part of the Fine-Grained Visual Categorization [FGVC9](https://sites.google.com/view/fgvc9) workshop at the Computer Vision and Pattern Recognition Conference [CVPR 2022](https://cvpr2022.thecvf.com/).\nA panel will review the top submissions for the competition based on the description of the methods provided.\nFrom this, a subset may be invited to present their results at the workshop.\nAttending the workshop is not required to participate in the competition; however, only teams that are attending the workshop will be considered to present their work.\n\nCVPR 2022 will take place virtually from June 19th to June 24th.\nPLEASE NOTE: CVPR frequently sells out early, we cannot guarantee CVPR registration after the competition's end.\nIf you are interested in attending, please plan ahead.\n\nYou can see a list of all of the FGVC9 competitions [here](https://sites.google.com/view/fgvc9/competitions).",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}