Leveraging Human Salience to Improve Calorie Estimation

TL;DR

This paper explores how integrating human salience maps can enhance calorie estimation from food images, achieving significant improvements but also revealing limitations in current models and pre-training strategies.

Contribution

It introduces the use of human salience maps for calorie prediction and evaluates their impact, highlighting challenges in surpassing existing benchmarks.

Findings

32.2% relative improvement with saliency maps

Pre-training on related tasks did not improve accuracy

Best models did not outperform original Nutrition5k results

Abstract

The following paper investigates the effectiveness of incorporating human salience into the task of calorie prediction from images of food. We observe a 32.2% relative improvement when incorporating saliency maps on the images of food highlighting the most calorie regions. We also attempt to further improve the accuracy by starting the best models using pre-trained weights on similar tasks of mass estimation and food classification. However, we observe no improvement. Surprisingly, we also find that our best model was not able to surpass the original performance published alongside the test dataset, Nutrition5k. We use ResNet50 and Xception as the base models for our experiment.