feat(jetson): Emotion detection — facial expression recognition for adaptive social responses #161

Closed
opened 2026-03-02 10:08:32 -05:00 by sl-jetson · 0 comments
Collaborator

Summary

Detect facial expressions to adapt robot conversation tone and responses.

Requirements

  • Expression classifier: 7 basic emotions (happy, sad, angry, surprised, fearful, disgusted, neutral)
  • Model: Lightweight CNN on TensorRT FP16 (face crop from SCRFD → expression classifier)
  • Per-person tracking: Track emotional state per person_id over time
  • Conversation adaptation: Feed emotion state to LLM context (e.g., be gentler if person seems sad)
  • ROS2 topics: /social/faces/expressions (Expression msg with person_id + emotion + confidence)
  • Privacy: Process on-device only, no cloud upload, optional opt-out per person
  • Performance: Piggyback on existing face detection pipeline, <5ms additional latency

Agent: sl-jetson

Labels: social-bot, emotion

## Summary Detect facial expressions to adapt robot conversation tone and responses. ## Requirements - **Expression classifier**: 7 basic emotions (happy, sad, angry, surprised, fearful, disgusted, neutral) - **Model**: Lightweight CNN on TensorRT FP16 (face crop from SCRFD → expression classifier) - **Per-person tracking**: Track emotional state per person_id over time - **Conversation adaptation**: Feed emotion state to LLM context (e.g., be gentler if person seems sad) - **ROS2 topics**: /social/faces/expressions (Expression msg with person_id + emotion + confidence) - **Privacy**: Process on-device only, no cloud upload, optional opt-out per person - **Performance**: Piggyback on existing face detection pipeline, <5ms additional latency ## Agent: sl-jetson ## Labels: social-bot, emotion
Sign in to join this conversation.
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: seb/saltylab-firmware#161
No description provided.