food_classifier

1.Business Problem:

1.1 Description

1. Dataset is taken from kaggle.
2. Dataset consist of different food listings of images. The dataset includes the set of images for each recipes.
3. Check Implementation from Kaggle

1.2 Problem Statemtent:

To build a CNN based model which can accurately detect food.

2.Dataset:

The dataset contains 5 sub-directories of food images.

• biryani
• burger
• dosa
• idly
• pizza

3.Buid CNN Model:

3.1 Model:

• Training accuracy is 88% and validation accuracy 80%.
• Model is fluctuating more.
• Let's use augumentation to increase data and check the result.

3.2 Augmentor:

Install Augmentor using command pip install Augmentor

• We can see that after using augumented data model is giving preety good accuracy.
• After using Augmentor accuracy has improved from 88% to 98%.
• Also loss and accuracy is not fluctuating.

4.Prediction: