Which facial recognition engines are supported?

It primarily focuses on the integration of DoubleTake with CompreFace for high-accuracy local facial embedding and matching.

Can I use this for wake word detection?

Yes, it includes deployment patterns for OpenWakeWord to enable custom wake word detection for voice satellites like Willow.

Does this skill work with Home Assistant?

Yes, it includes detailed configurations for integrating the Wyoming voice pipeline and MQTT event data directly into Home Assistant's Assist Pipeline.

What hardware acceleration does this skill support?

The skill provides specific configurations for AMD GPUs using ROCm and OpenVINO, as well as NVIDIA GPUs using CUDA for Whisper and object detection.

Is the voice processing local or cloud-based?

The perception pipeline is designed for 100% local processing using Whisper for speech-to-text and Piper for text-to-speech, ensuring maximum privacy.

Perception Pipeline

Name: Perception Pipeline
Author: Lobbi-Docs

byLobbi-Docs

Ciencia de Datos y ML

Deploys and manages advanced smart home perception layers including NVR, facial recognition, and local voice processing pipelines.

Acerca de

The Perception Pipeline skill provides specialized knowledge for building sovereign, local-first smart home environments using the Ahling Command Center architecture. It offers domain-specific guidance for configuring Frigate NVR with AMD ROCm hardware acceleration, implementing facial recognition via DoubleTake and CompreFace, and orchestrating private voice pipelines using Whisper, Piper, and OpenWakeWord. This skill is ideal for developers seeking to automate complex video and audio event processing while maintaining data privacy and high-performance local execution.

Características Principales

Implements local facial recognition using DoubleTake and CompreFace
Configures Frigate NVR with AMD GPU (ROCm) hardware acceleration
Sets up Wyoming-compatible voice pipelines for STT and TTS
Manages MQTT-based event structures for real-time automation
Provides Docker Compose templates for perception-layer services
0 GitHub stars

Casos de Uso

Integrating privacy-focused voice satellites with Home Assistant
Automating home security through facial recognition and custom MQTT events
Building a secure, local-only video surveillance system with object detection

Acerca de

Características Principales

Implements local facial recognition using DoubleTake and CompreFace
Configures Frigate NVR with AMD GPU (ROCm) hardware acceleration
Sets up Wyoming-compatible voice pipelines for STT and TTS
Manages MQTT-based event structures for real-time automation
Provides Docker Compose templates for perception-layer services
0 GitHub stars

Casos de Uso

Integrating privacy-focused voice satellites with Home Assistant
Automating home security through facial recognition and custom MQTT events
Building a secure, local-only video surveillance system with object detection