The aim of the project was to classify food image datasets afforded by the Snap Food application.
Sellers can put their food items on Snap Food so that users can choose the food they want by checking the price and photos of the food and the comments of other users.
By checking the pictures that people share on social networks, you can improve the arrangement of food. For example, put the foods that have been noticed in social networks at the beginning of the site or application. It also suggested a personalized list of his favorite foods for each user.
To use these facilities, in the first step, we must be able to identify the type of food in a photo, and this is exactly what I do in this project.
The dataset contains 21 categories of food images was been searched by users in social media. In this collection, there are more than 17,000 photos of 21 types of food, some of which you can see in the above photo:
- Number of categories: 21
- Number of images: 17552
For implementation I use pre-trained vision models to do the classification.
