Food Image Processing Using Deep Learning

Image Processing, Artificial Intelligence

Healthy lifestyle is an important factor for disease free survival in this pandemic era. To maintain proper health, it is essential to consume good nutritious food. This research embarks on a pivotal journey to harness the capabilities of automatic food image recognition systems, with a profound focus on nutrient identification. Such a system stands poised to usher in a transformative wave within the domains of computer vision, dietary analysis, and fitness monitoring. The objective of this study is to train a deep learning model which identifies a food image captured with any camera device and generate a nutrient estimate report. The proposed model is a two-step process, firstly it recognizes a processed food image using a deep learning model, then it generates a dietary assessment report based on a synthesized nutrient value dataset taking USDA national nutrient database as a standard source. Transfer learning is applied on VGG16 model architecture and experimental evaluation is conducted using subset of food image dataset FOOD-101. Remarkable outcomes emerge through the adaptation of a new CNNovaNet architecture, yielding exceptional results, particularly when applied to the complete FOOD-101 dataset.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.