Video Object Detection

Install

$ pip install -r requirements.txt

Additionally requires:

• OpenCV 3.4+
• Movidius NCSSDK v2
• Tensorflow (optional)

Setup

1. Install opencv.
2. Add models to ./models/<model_dir>
3. Add ./CONFIG file; see ./CONFIG_EXAMPLE for reference.

Usage

Monitor Mode

Inference Mode

$ python -m src.video -i 0 -c CONFIG

Test Mode

$ python -m src.video -i 0 -o tmp/test.mov -c CONFIG