Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Kitchen Food Waste Image Segmentation and Classification for Compost Nutrients Estimation

Raiyan Rahman, Mohsena Chowdhury, Yueyang Tang, Huayi Gao, George Yin, Guanghui Wang

TL;DR

This work addresses estimating compost nutrients (NPK) from kitchen waste by applying semantic segmentation to a newly collected high-resolution image dataset of 19 nutrition-rich waste classes. It benchmarks four state-of-the-art segmentation models, finding SegFormer with the MIT-B5 backbone to achieve the highest mean IoU (~67.08) across 10-fold cross-validation. The dataset comprises 2,912 curated images with 29,433 annotations, derived from an initial 3,128 high-resolution kitchen-waste images and refined to focus on nutritionally relevant classes. The findings advocate transformer-based approaches for accurate waste-class delineation and set the stage for automatic NPK estimation from segmented waste to guide composting practices.

Abstract

The escalating global concern over extensive food wastage necessitates innovative solutions to foster a net-zero lifestyle and reduce emissions. The LILA home composter presents a convenient means of recycling kitchen scraps and daily food waste into nutrient-rich, high-quality compost. To capture the nutritional information of the produced compost, we have created and annotated a large high-resolution image dataset of kitchen food waste with segmentation masks of 19 nutrition-rich categories. Leveraging this dataset, we benchmarked four state-of-the-art semantic segmentation models on food waste segmentation, contributing to the assessment of compost quality of Nitrogen, Phosphorus, or Potassium. The experiments demonstrate promising results of using segmentation models to discern food waste produced in our daily lives. Based on the experiments, SegFormer, utilizing MIT-B5 backbone, yields the best performance with a mean Intersection over Union (mIoU) of 67.09. Class-based results are also provided to facilitate further analysis of different food waste classes.

Kitchen Food Waste Image Segmentation and Classification for Compost Nutrients Estimation

TL;DR

This work addresses estimating compost nutrients (NPK) from kitchen waste by applying semantic segmentation to a newly collected high-resolution image dataset of 19 nutrition-rich waste classes. It benchmarks four state-of-the-art segmentation models, finding SegFormer with the MIT-B5 backbone to achieve the highest mean IoU (~67.08) across 10-fold cross-validation. The dataset comprises 2,912 curated images with 29,433 annotations, derived from an initial 3,128 high-resolution kitchen-waste images and refined to focus on nutritionally relevant classes. The findings advocate transformer-based approaches for accurate waste-class delineation and set the stage for automatic NPK estimation from segmented waste to guide composting practices.

Abstract

The escalating global concern over extensive food wastage necessitates innovative solutions to foster a net-zero lifestyle and reduce emissions. The LILA home composter presents a convenient means of recycling kitchen scraps and daily food waste into nutrient-rich, high-quality compost. To capture the nutritional information of the produced compost, we have created and annotated a large high-resolution image dataset of kitchen food waste with segmentation masks of 19 nutrition-rich categories. Leveraging this dataset, we benchmarked four state-of-the-art semantic segmentation models on food waste segmentation, contributing to the assessment of compost quality of Nitrogen, Phosphorus, or Potassium. The experiments demonstrate promising results of using segmentation models to discern food waste produced in our daily lives. Based on the experiments, SegFormer, utilizing MIT-B5 backbone, yields the best performance with a mean Intersection over Union (mIoU) of 67.09. Class-based results are also provided to facilitate further analysis of different food waste classes.
Paper Structure (7 sections, 2 equations, 2 figures, 3 tables)

This paper contains 7 sections, 2 equations, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Dataset
  4. Models
  5. Experiments
  6. Results and Analysis
  7. Conclusion

Figures (2)

  • Figure 1: The home composter would conveniently allow households to recycle their kitchen scraps and daily food waste into high-quality compost rich in nutrients. This paper aims to study the feasibility of using semantic segmentation techniques to identify food waste classes, enabling the capture of nutritional information from food waste.
  • Figure 2: Visual results comparing the ground truths to their corresponding segmentation predictions are presented for cases involving egg shells, onion skins, and hard bread across various segmentation models.