Guiding Philosophy:
Food for Thought, Thought for Food

  • Ashoka University, India

FKG.in Dataset @ Blazegraph

Coming Soon!

RASOI Dataset @ Kaggle

The RASOI (Recognition and Segmentation of Indian Thali using AI) dataset is focused on Indian cuisine images and its annotations. It aims to advance AI-driven food analysis by focusing on the visually complex and culturally diverse Indian Thali images. This dataset was created for the RASOI challenge, organized as part of NCVPRIPG 2025 conference. It comprises images of Indian dishes, categorized into:
  • Standalone Recipes: Individual food items with corresponding labels.
  • Composite Thalis: Images containing multiple dishes with bounding box annotations and labels.
Kaggle Link: RASOI dataset

License: This dataset is provided under the CC BY-NC-SA 4.0 license and is intended solely for non-commercial research use.

Attribution Requirement: Any use of this dataset must include proper credit to its creators. The following attribution must appear in all publications, presentations, evaluations, communications, and anywhere else that the dataset may be referenced:
"Created and maintained by the Food Computing Lab, Ashoka University, India."