qualcomm-ultralytics-ppe_detection

• Model creator: Ultralytics
• Original model: YOLOv11n

This repository contains the YOLOv11n model optimized for Qualcomm hardware using Qualcomm® AI Engine Direct (SNPE).

It is designed for high-performance, real-time object detection inference on edge devices powered by Qualcomm Snapdragon platforms, enabling efficient on-device AI capabilities with low latency and reduced power consumption.

Model Details

• Developed by: Advantech-EIOT / Ultralytics
• Architecture: YOLOv11n
• Task: PPE Detection
• Precision: Quantized (w8a16) for NPU optimization
• Input Resolution: 160 x 160
• Calibration Data: Calibrated using 1,000 images
• Optimization: Qualcomm® AI Stack / SNPE SDK

Hardware Compatibility

This model is highly optimized for Advantech Edge AI platforms powered by Qualcomm processors:

• Linux: Dragonwing® Platforms (e.g. Dragonwing® IQ-9075)

Limitations and Disclaimer

YOLOv11 is a powerful real-time object detection model, but it may exhibit limitations depending on the environment and deployment context.

• Accuracy: The model's accuracy, especially for small objects or in low-light conditions, may be impacted by the reduced input resolution (160x160) and quantization (w8a16). Users should validate outputs for critical applications or safety-critical systems.
• Usage: Please refer to the Ultralytics AGPL-3.0 License for usage restrictions and acceptable use policies.
• Edge Optimization: Inference performance (FPS) and bounding box precision may vary depending on the specific hardware configuration, camera ISP pipelines, and thermal constraints of the edge device.