Live Projects

Project Overview

• What I’m building: I’m building a desktop system that monitors my face via webcam for fatigue and distraction.
• Why I’m building it: I want a practical tool to help me maintain focus during long work sessions.
• Why it matters: It can improve my productivity and wellbeing by giving me real-time feedback.

Current Progress & Methods

• I have finished developing the core system components. This includes the webcam capture, facial feature extraction, face-tracking servo control, and the real-time HUD display. I also completed the data labeling pipeline and trained the ensemble fatigue detection models. The final step of deploying the inference engine, which uses these models to trigger alerts, is now done.
• I am integrating the newly deployed inference engine into the main capture loop. I am running end-to-end tests to confirm real-time performance and alert accuracy.

# From capture_and_extract.py (conceptual main loop logic)
from hud_overlay import HUD
from servo_controller import ServoController
from alert_engine import AlertEngine # Used for inference

# ... setup for HUD and ServoController ...
hud = HUD()
hud.start()
servo = ServoController()
alert_engine = AlertEngine(servo=servo)

# In the continuous frame processing loop:
# ... after feature_dict is computed for the current frame ...
state, probabilities = alert_engine.push_frame(feature_dict)
hud.update(state, my_metrics_dict) # my_metrics_dict includes EAR, MAR, etc.

Visuals