2018 FGVCx Fungi Classification Challenge

The 2018 competition is part of the FGVC^5 workshop at CVPR. Our sponsor, the Svampe Atlas, has provided a dataset from a carefully curated database containing over 100,000 fungi images. The Svampe Atlas has a comprehensive representation of nearly 1,500 wild mushrooms species which have been spotted and photographed by the general public in Denmark.

Please open an issue if you have questions or problems with the dataset.

Kaggle

We are using Kaggle to host the leaderboard. Checkout the competition page here.

Dates


Data Released	April 1, 2018
Submission Deadline	June 11, 2018
Winners Announced	June 22, 2018

Details

There are a total of 1,394 fungi species in the dataset, with 85,578 training images and 4,182 validation images. The testing set contains 9,758 images. All images are sourced from fungi species submitted in the Danish Svampe Atlas and example images can be viewed here.

Evaluation

We follow a similar metric to the classification tasks of the ILSVRC. For each image , an algorithm will produce 3 labels , . For this competition each image has one ground truth label , and the error for that image is:

Where

The overall error score for an algorithm is the average error over all test images:

Guidelines

Participants are restricted to train their algorithms on the 2018 FGVCx Fungi Classification competition train and validation sets. Pretrained models may be used to construct the algorithms (e.g. ImageNet pretrained models) as long as participants do not actively collect additional data for the target species in the 2018 FGVCx Fungi Classification competition. Please specify any and all external data used for training when uploading results.

The general rule is that we want participants to use only the provided training and validation images to train a model to classify the test images. We do not want participants crawling the web in search of additional data for the target categories. Participants should be in the mindset that this is the only data available for those categories.

Participants are allowed to collect additional annotations (e.g. bounding boxes) on the provided training and validation sets. Teams should specify that they collected additional annotations when submitting results.

Annotation Format

We closely follow the annotation format of the COCO dataset. The annotations are stored in the JSON format and are organized as follows:

{
  "info" : info,
  "images" : [image],
  "categories" : [category],
  "annotations" : [annotation],
  "licenses" : [license]
}

info{
  "year" : int,
  "version" : str,
  "description" : str,
  "contributor" : str,
  "url" : str,
  "date_created" : datetime,
}

image{
  "id" : int,
  "width" : int,
  "height" : int,
  "file_name" : str,
  "license" : int,
  "rights_holder" : str
}

category{
  "id" : int,
  "name" : str,
  "supercategory" : str,
}

annotation{
  "id" : int,
  "image_id" : int,
  "category_id" : int
}

license{
  "id" : int,
  "name" : str,
  "url" : str
}

Submission Format

The submission format for the Kaggle competition is a csv file with the following format:

id,predicted
12345, 23 3 42
67890, 42 0 21

The id column corresponds to the test image id. The predicted column corresponds to 3 category ids, separated by spaces. You should have one row for each test image.

Terms of Use

By downloading this dataset you agree to the following terms:

You will abide by the Danish Svampe Atlas Terms of Service
You will use the data only for non-commercial research and educational purposes.
You will NOT distribute the above images.
The Danish Svampe Atlas makes no representations or warranties regarding the data, including but not limited to warranties of non-infringement or fitness for a particular purpose.
You accept full responsibility for your use of the data and shall defend and indemnify the Danish Svampe Atlas, including its employees, officers and agents, against any and all claims arising from your use of the data, including but not limited to your use of any copies of copyrighted images that you may create from the data.

Data

Download the dataset files here:

Training and validation images [13GB]
- Running md5sum fungi_train_val.tgz on the tgz file should produce df2b2980835668ed1d9a0e286e4bdadd
- Images have a max dimension of 1024px and have been converted to JPEG format
- Untaring the images creates a directory structure like images/category/image.jpg. This may take a while.
Testing images [1.3GB]
- Running md5sum fungi_test.tgz on the tgz file should produce 949fc7266f7c6574e4ce359cf2571c85
- Images have a max dimension of 1024px and have been converted to JPEG format
- Untaring the images creates a directory structure like test/image.jpg. This may take a while.
Training and validation annotations [2.9MB]
- Running md5sum train_val_annotations.tgz on the tgz file should produce `c3c0f1b8a8d0b60619c43d9c89bdbc7e'
Testing image information [92KB]
- Running md5sum test_information.tgz on the tgz file should produce 2f8a8f361e59d27d9ce4c96f4dec3817

axelsparr/fgvcx_fungi_comp