Important Notice
This model is provided for evaluation and development purposes only. It is not validated for and must not be used in clinical, diagnostic, or production settings. See #use-and-limitations.
Model Information
Description
people-present is a single-stage object detection model that detects caretakers and other persons in a NICU warmer video feed. It uses the same MobileNetV2 + FPN + anchor-free head architecture as the companion patient-present model, retrained on a person-detection objective. In the NICU Warmer reference pipeline, its bounding boxes are cross-referenced with patient-present detections via IoU overlap to classify each detected person as either "caretaker" or "patient". This model is a development and evaluation tool produced as part of Intel's NICU Warmer reference design and has not undergone clinical validation.
Intended Use
This model is intended for use by software developers and researchers evaluating AI-assisted NICU monitoring workflows on Intel hardware. It demonstrates person detection in warmer video as a building block for caretaker-presence sensing. people-present itself does not provide any medical functionality, nor is it intended to process or interpret medical data for a medical purpose. Developers are responsible for independently validating and adapting people-present for their specific use case. It must not be used in live clinical environments or for any diagnostic or patient-safety purpose.
Technical Specifications
| Attribute | Detail |
|---|---|
| Architecture | MobileNetV2 backbone + FPN neck + anchor-free detection head |
| Parameters | ~5–7M (FP32) |
| Input | 992×800 RGB image (W×H), float32, normalized [0,1], NCHW |
| Output | [N, 5] — bounding boxes (x1, y1, x2, y2, confidence), post-NMS, pixel coords at input resolution |
| Training hardware | Intel Ultra Core |
| Framework | PyTorch (mmdetection) → OpenVINO IR FP32 |
Use and Limitations
Permitted Uses
- Evaluation and benchmarking of [domain] AI workflows
- Research and development
- Academic study
- Adaptation and modification for non-clinical applications
Prohibited Uses
- Clinical or diagnostic use
- Production deployment without independent validation and applicable regulatory authorization
- Use with live patients or real patient data (unless the user has obtained all necessary authorizations)
- Any use that would cause Intel to be treated as the manufacturer, provider, or deployer of a medical device or regulated AI system
License
The use of people-present is governed by the Intel Non-Clinical Evaluation Use License. By accessing this model, you agree to the license terms.