Standard Bank Tech Impact Challenge: Animal classification
Knowledge
Create a binary classification algorithm to distinguish zebras from elephants
241 data scientists enrolled, 57 on the leaderboard
ConservationComputer VisionGood for BeginnersUnstructuredImage
30 August 2019

The data have been split into a test and training set. The training set contains 13,999 images of animals and the test set contains 5000 images. There are two types of animals in this dataset, zebras and elephants.

Data was retrieved from the Data Repository for the University of Minnesota, https://doi.org/10.13020/D6T11K, under a creative commons license, from a study titled: Camera Trap Images used in "Identifying Animal Species in Camera Trap Images using Deep Learning and Citizen Science". All images were downloaded from Zooniverse and have been resized to 330x330 pixels.

“The cameras were located in Tanzania, at various places in the Serengeti National Park. The cameras were triggered by a combination of infrared and motion sensors and took three images, after which the trigger was blocked for one minute.” For more information on the dataset, please visit https://www.snapshotserengeti.org/ and https://conservancy.umn.edu/handle/11299/199819.

Your task is to provide the probability that an image contains a zebra. For each unique image ID you should estimate the likelihood that the image contains a zebra, with an estimated probability value between 0 and 1.

Files available for download

The files you have for download here are:

  • train_elephants.zip (~137mb): zip file containing all the images of elephants in the training set. You will use these images to train your model.
  • train_zebras.zip (~157mb): zip file containing all the images of zebra in the training set. You will use these images to train your model.
  • test.zip (~126mb): zip file containing all the images of elephants and zebra in the test set. You will use these images to test your model on.
  • sample_submission.csv: is an example of what your submission file should look like, including a list of unique image IDs. Your submission file should have all of the Image IDs in this file along with estimates of the corresponding probabilities of observing a zebra.

Turn your model into a web app with this blog by Johnowhitaker.