Ingredient segmentation with transparency

In the food industry, the demand for the extraction of food ingredients from images is increasing for quality control and appearance evaluation of products in the manufacturing field. Therefore, recognition of food ingredients distribution using deep learning has been proposed. However, this method still has a problem of misrecognizing transparent ingredients as other ingredients. To solve this problem, we propose a method for generating training data that includes the transparency of food ingredients in order to improve the accuracy of food ingredient identification.