Skip to content

Framework Integration

PrimateFace provides integration with popular computer vision and pose estimation frameworks.

Overview

These guides help you use PrimateFace with external frameworks:

  • MMDetection/MMPose - Primary framework for detection and pose estimation
  • DeepLabCut - Alternative pose estimation with COCO training support
  • SLEAP - Multi-animal tracking with COCO training support
  • YOLO - Real-time detection and integration examples

Integration Approach

PrimateFace doesn't replace these frameworks - it provides:

  1. Training Scripts - Convert COCO data to framework-specific formats
  2. Model Compatibility - Use framework models with PrimateFace pipelines
  3. Evaluation Tools - Compare performance across frameworks
  4. Workflow Integration - Seamless integration with existing workflows

When to Use Each Framework

  • MMDetection/MMPose - Production inference, best performance
  • DeepLabCut - Markerless tracking, behavioral analysis
  • SLEAP - Multi-animal scenarios, complex tracking
  • YOLO - Real-time applications, edge deployment

Getting Started

Each framework has specific setup requirements. See the individual guides for:

  • Installation and setup
  • Training from COCO data
  • Integration with PrimateFace
  • Performance optimization

For detailed API documentation, see the API Reference.