Instructions for Running the Code for Eatalyze: Using Machine Learning to Classify and Analyze Food Nutrition

Authors: Elisabeth Kollrack, Ellie Sheppard, Aaron Wilson

Eatalyze is a machine learning project that classifies foods and analyzes their nutritional properties using traditional machine learning models and deep learning techniques. It includes:

• Nutritional analysis using Linear Regression and XGBoost
• A custom-trained image classifier using the Food-101 dataset
• A live webcam-based food classifier that predicts food categories in real time

Linear Regression and XGBOOST Scripts (Located in the Code/ Folder)

pip install the following packages:

1. pip install numpy
   

2. pip install pandas
   

3. pip install matplotlib
   

4. pip install scikit-learn
   

5. pip install statsmodels
   

6. pip install xgboost
   

Ensure the food.xlsx data is installed and in the same folder as your jupyter notebooks for this line to run food = pd.read_excel("food.xlsx", sheet_name="food"). From there, just run all cells and the graphs and tables should appear!

ML_Custom_Classifier.py

Below are instructions to run the training script, which will train the model and save the parameters to a file on your computer. This section will guide you on how to download the food-101 images and install the required packages.

pip install the following packages:

1. pip install numpy
   

2. pip install pandas
   

3. pip install matplotlib
   

4. pip install scikit-learn
   

5. pip install tensorflow
   

To download the food-101 dataset:

import kagglehub
path = kagglehub.dataset_download("kmader/food41")
print("Path to dataset files:", path)

Now run the program:

• Update the dataset path in the program to the path on your specific computer
• Run this updated script and ensure that the Epochs are incrementing
• Watch grass grow : )

ML_Image_Classifier.py

The ML_Image_Classifier.py script opens a new window and uses the computer camera to classify objects in front of it. This uses the model that was previously trained with ML_Custom_Classifier.py. This will also require a few packages to run.

pip install the following packages:

1. pip install numpy
   

2. pip install pandas
   

3. pip install matplotlib
   

4. pip install scikit-learn
   

5. pip install tensorflow
   

6. pip install opencv-python
   

Ensure the model is correct:

model = tf.keras.models.load_model("food_cnn_model_custom.h5")

Now run the program:

• A new window will open using the built-in laptop camera
• When the desired food is within this frame, press the "s" key to capture an image
• This image will then be passed through the classifier to predict the image's label
• The resulting prediction will output to the terminal with an associated confidence level
• Press the "q" key to quit the program and exit