Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Revamping Cross-Modal Recipe Retrieval with Hierarchical Transformers and Self-supervised Learning

About

Cross-modal recipe retrieval has recently gained substantial attention due to the importance of food in people's lives, as well as the availability of vast amounts of digital cooking recipes and food images to train machine learning models. In this work, we revisit existing approaches for cross-modal recipe retrieval and propose a simplified end-to-end model based on well established and high performing encoders for text and images. We introduce a hierarchical recipe Transformer which attentively encodes individual recipe components (titles, ingredients and instructions). Further, we propose a self-supervised loss function computed on top of pairs of individual recipe components, which is able to leverage semantic relationships within recipes, and enables training using both image-recipe and recipe-only samples. We conduct a thorough analysis and ablation studies to validate our design choices. As a result, our proposed method achieves state-of-the-art performance in the cross-modal recipe retrieval task on the Recipe1M dataset. We make code and models publicly available.

Amaia Salvador, Erhan Gundogdu, Loris Bazzani, Michael Donoser• 2021

Related benchmarks

TaskDatasetResultRank
Image ClassificationFood-101 (test)
Accuracy84.44
145
Image-to-recipe retrievalRecipe1M 10k setup (test)
Recall@133.5
125
Recipe-to-image retrievalRecipe1M 10k setup (test)
R@133.7
120
Image-to-recipe retrievalRecipe1M 1k setup (test)
Recall@164.2
116
Recipe-to-image retrievalRecipe1M 1k setup (test)
Recall@164.5
110
Image-to-recipe retrievalRecipe1M 1.0 (test)
Median Rank1
35
Recipe-to-image retrievalRecipe1M 1.0 (test)
MedR1
30
Cross-modal Retrieval (Image-to-Recipe)Recipe1M v1 (1k)
MedR1
28
Food RecognitionISIA Food-500 (test)
Accuracy57.562
14
Cross-modal Retrieval (Recipe-to-Image)Recipe1M v1 (1k)
Median Rank1
13
Showing 10 of 11 rows

Other info

Code

Follow for update