Back to projects

On this page

  1. Real-time Dog and Cat Classifier
  2. Project overview
  3. Proposed solution
  4. Technologies used
  5. Main features
  6. Results and project value

Real-time Dog and Cat Classifier

Intelligent system for identifying dogs and cats in images and live video using convolutional neural networks.

October 10, 2023
JavaScriptPythonHTMLCSSCNNTensorFlow

Real-time Dog and Cat Classifier

Classifier Banner

Project overview

This project addressed the need to automate visual animal identification through artificial intelligence. In contexts such as animal shelters, smart surveillance systems, or educational tools, it is essential to quickly determine whether an image or live video corresponds to a dog or a cat. Manual classification is slow, error-prone, and difficult to scale when dealing with large volumes of data.

Proposed solution

A real-time classification system was developed based on convolutional neural networks (CNNs), trained with TensorFlow and Python. The application can process static images and live webcam input, identifying in each frame whether the subject is a dog or a cat with minimal latency.

The project included dataset preparation, model training, validation, and export for integration into a lightweight web interface built with JavaScript, HTML, and CSS. Thanks to this integration, predictions run directly in the browser without requiring a complex backend inference process.

Technologies used

  • Python for data preparation, model training, and export
  • TensorFlow for building and training the CNN
  • Convolutional neural networks for visual feature extraction
  • HTML and CSS for the user interface
  • JavaScript for camera interaction and web experience

Main features

  • Real-time classification from webcam
  • Lightweight, responsive, and user-friendly interface
  • Custom-trained model for two classes
  • Full integration between machine learning and user interaction

Results and project value

The development demonstrates how an artificial intelligence model can move from an experimental environment to a practical and accessible solution for end users. It also illustrates the direct application of TensorFlow and Python in real-time inference systems, reinforcing the connection between applied research and interactive digital products.

Back to projects