Skip to content

Latest commit

 

History

History
33 lines (23 loc) · 1.82 KB

README.md

File metadata and controls

33 lines (23 loc) · 1.82 KB
Raw

Smoker Detection Using Convolutional Neural Networks

This project is an exploration into the capabilities of Convolutional Neural Networks (CNNs) in classifying images to detect smoking behavior. The aim is to demonstrate how machine learning can be applied to a challenging and nuanced task like smoker detection.

About the Project

This project uses a CNN to predict whether a person in a photo is smoking. The model was trained on a diverse dataset and went through several iterations to improve its accuracy and reduce overfitting. The journey from the initial complex model to the final refined version is a tale of learning and adapting in the field of AI and machine learning.

Key Features:

  • Utilizes TensorFlow and Keras for building and training the CNN model.
  • Employs techniques like one-hot encoding, hyperparameter tuning, and callbacks like Early Stopping and Model Checkpoints.
  • Integrates Weights and Biases for experiment tracking and performance visualization.
  • Contains detailed documentation and reflections on the model's performance and future improvement areas.

Getting Started

To get started with this project, you'll want to clone the repository and explore the Smoker_detection_final.ipynb Jupyter notebook. This notebook contains the complete code, from data preprocessing to model training and evaluation.

Prerequisites:

  • Python 3.x
  • TensorFlow 2.x
  • Other Python libraries as specified in the notebook

Installation:

Clone the repository using:

git clone https://github.com/markredito/Smoker_Detection.git

Navigate to the cloned directory and open the Jupyter notebook to get started.

Dataset

The dataset used in this project can be found here. It's a well-curated collection of images labeled as 'Smoking' and 'Not Smoking'.