sightsprite 0.1.1

Machine vision with voice alerts

sightsprite

Real-time machine vision with voice alerts.

Fun with machine vision. Utilities for all stages of pipeline development for machine vision with voice alerts: from data collection to real-time inference.

Under rapid and active development, breaking changes are guaranteed through summer of 2025. Am currently building:

• OpenCV utilities for capturing training data.
• Simple tools (in matplotlib) for labeling and organizing captured data.
• Tools for training pytorch models.
• Utilities for realtime inference with voice alerts (e.g., "get off the couch").

Everything is changing very fast. Will update when things are less unstable.

To do

Once following steps working I'll consider sightsprite to be in alpha stage:

1. Get capture.py in slightly better shape
   • Add ability to change width/height instead of default width x height
   • fix up the weird Qt warnings in linux
   • Make some simple tests
2. Train on sleep data (use deepglue), build out training.py module.
   • Label sleep data using ImageLabeler
   • Train network using deepglue tools within sandbox.
3. Once you have model trained, build out out inference.py
   • Real-time inference
   • Voice alerts to nudge you.
4. Build some docs to explain how it all works.

Notes etc

Any images skipped during labeling will appear first when you restart the ImageLabeler