Food Multiclassification 🥗

The classificier predicts an array of labels for a single image of food. Trained on images of 101 different dishes, such as apple pie, greek salad, sushi, etc.

Using:

• Tensorflow 1.15
• Python 3.7
• Sklearn
• Pandas
• OpenCV

Predicting models:

• Xception
• ResNet50
• DenseNet121

---

The labels:

• healthy
• junks
• desserts
• appetizers
• mains
• soups
• carbs
• proteins
• fats
• meat

---

Usage:

Create a folder. Then using your console write git clone https://github.com/tableClothed/food--multiclassification.git.

When the repo finishes cloning, enter the food--multiclassification folder.

Using your Anaconda console in here use jupyter anaconda command.

In your working tree click on models.ipynb and train whichever predicting model you like (Xception, ResNet50 or DenseNet121).

After that enter the models.ipynb file, find model = tensorflow.kerasmodels.load_model({MODEL}) and in place of {MODEL} insert the name of model, you'd like to use for making predictions.

If you wish to add more testing images, you can paste them to 'data/test' folder.