ICLR Workshop Challenge #1: CGIAR Computer Vision for Crop Disease
$7,000 USD
Identify wheat rust in images from Ethiopia and Tanzania, and win a trip to present your work at ICLR 2020 in Addis Ababa.
29 January–15 March 2020 23:59
566 data scientists enrolled, 195 on the leaderboard
published 30 Jan 2020, 11:25

any one to provide a starter notebook for a novice


Here's one I just threw together. Let me know if you have any questions. You can also look around online for image classification tutorials - many will have the data in the same format as we use here - a folder for each class and optionally a test folder of images to classify. It's the same approach for dogs vs cats as for rust vs healthy :) There are many libraries and models to choose from. Good luck :)

Hey John, i haven't decided to join this one yet. What's the image format ? How many pixels?

Also, did you reach your current score with fastai only?

Images are varied - some low res (~400x600), some higher resolution (~3000x2000). Current score is with fastai, nothing too crazy or custom :) I think there's potential for some good tricks with data augmentation, especially cropping into some of the higher-res images where the rust is only on a few plants.

Hi John, is there a way to make Fast AI read the .jfif files. PIL.Image.register_extension = '.jfif' isn't working.

You can easily convert them to jpg images

Hello Stanley,

Please is it possible for us to talk privately, I would appreciate if we can discuss some things.

Please share your LinkedIn profile or any other social media or means you feel I can reach you on