Dog Barking DetectorThis is a dog barking detection tool. Completely open source and runs offline. It uses machine learning recognition to log when a dog barks. Useful for keeping track of when your dog is barking!
This tool uses the YAMNet audio classification model, a deep neural network trained on the AudioSet dataset. It processes audio in real-time, analyzing 1-second chunks at 16kHz sample rate. The model classifies sounds into 521 categories, with specific focus on dog vocalizations (classes 70-75: Bark, Yip, Howl, Bow-wow, Growling, Whimper, etc).
Built with TensorFlow.js for client-side inference, ensuring complete privacy as all processing happens in your browser. Audio capture and processing is handled through the Web Audio API.
The YAMNet model is loaded from TensorFlow Hub and converted to TensorFlow.js format for browser compatibility.
This project is completely open source and runs entirely without a server. No data is sent to any servers or 3rd party services - everything happens in your browser.
Built with modern web technologies: HTML5, CSS3, JavaScript, TensorFlow.js
License: MIT
Contact me with any questions, feedback, bugs or feature requests. Please consider supporting the project on github as well.
I build small side projects, plugins, chrome extensions, and digital art - always tinkering. Contact me at https://j7digital.com/
| Timestamp | Sound Recorded | Play |
|---|
