Kitchen Food Waste Image Segmentation and Classification for Compost Nutrients Estimation
This work addresses food waste management for households by providing a tool to assess compost quality, but it is incremental as it applies existing models to a new dataset.
The paper tackled the problem of estimating compost nutrients from kitchen food waste by creating a new annotated image dataset and benchmarking four semantic segmentation models, with SegFormer achieving the best performance at 67.09 mIoU.
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.