Invitation to Food Recognition Benchmark 2022

3 views

Skip to first unread message

Shivam Khandelwal

unread,

Mar 16, 2022, 3:06:42 PM3/16/22

to Open Food

We are happy to invite you to the Food Recognition Benchmark 2022.

# Goal

The goal of this benchmark is to train models which can look at images of food items and detect the individual food items present in them.

# Dataset

## What’s different about it?

To put it bluntly: most food images on the internet are a lie, but algorithms need to work on real-world images.

Search for any dish, and you’ll find beautiful stock photography of that particular dish. It is same on social media: we share photos of dishes with our friends when the image is exceptionally beautiful.

We use a novel dataset of food images collected through the MyFoodRepo app, where numerous volunteer Swiss users provide images of their daily food intake in the context of a digital cohort called Food & You.

This growing data set has been annotated - or automatic annotations have been verified - with respect to segmentation, classification (mapping the individual food items), and weight/volume estimation. This is an evolving dataset, where we will continue to release more data as the dataset grows over time.

## Dataset Size

Training set

- 54392 images of food items,

- 100,256 annotations,

- 323 food classes.

Testing set

- 3,000+ unreleased images

Dataset Preview: https://i.imgur.com/BjH4ypx.png

# Prizes and Authorship

Top participants of the Benchmark will be invited to be co-authors of the dataset release paper and the challenge solution paper.

- 1st Prize: DJI FPV Drone Combo

- 2nd Prize: DJI Mavic Air 2

- 3rd Prize: Oculus Quest 2

# Deadline: May 3rd, 2022

# Baselines and Starter Kit

This benchmark comes with Detectron2 and MMdetection implementation on the dataset in getting started baseline, so you don’t need to worry about the initial implementation barrier.

# Important Links

Benchmark: https://www.aicrowd.com/challenges/food-recognition-benchmark-2022

Benchmark Leaderboard: https://www.aicrowd.com/challenges/food-recognition-benchmark-2022/leaderboards

Detectron2 & MMdetection Baseline https://discourse.aicrowd.com/t/resources-for-mmdetection-submissions/7329
https://discourse.aicrowd.com/t/making-detectron2-submissions-is-now-even-easier/7195

Dataset Download: https://www.aicrowd.com/challenges/food-recognition-benchmark-2022/dataset_files