[Week6 — Eat & Count]
In our project, our goal is to understand a given photos’ class and calculate the calorie of foods eaten in a period. In previous blogs, we talked about image classification, how to understand the kind of a given test food photo. In this blog, we will be talking about the second part of our project, how we get the calories of foods and how we extract our second data set.
For the calories of foods, we cannot find a data set that satisfied our need for our project. Then we decided to extract our own data set. After classifying the test photo, we did not measure the size of food, we assume it like one serving. Therefore, in our data set, we have got one serving calories of foods. As a future work, we will improve our project by understanding the size of foods and calculate the calories accordingly the size.
In order to find true values of calories, we find the calorie values from at least three different resources and compare them to find the true values. If we could not find the values of one serving food, we find 100g values and we calculated the calories per serving.
After collecting all values for out Food-101 dataset, we saved them and we will use the values after classifying a given photo. On next blogs, we will talk about pre-trained neural networks and the result of our experiments with it. See you on next blogs.
References
