This Python app allows user to parse MCAP files and view each H264 frame using Matplotlib. It also allows the user to overlay the bounding boxes from the detection topic (in blue) onto the frame. Furthermore, the app also has a feature to feed it an Onnx model which can inference the frame and overlay those bounding boxes (in red) onto the frame. Additional support and documentation for the application can be found on here.

The Tkinter library must be installed as part of the Ubuntu environment using the following command:

sudo install python3-tk

Python library requirements must be installed via Pip:

pip install -r requirements.txt

The script can be run simply with:

python viewer.py <mcap_file>

To view the MCAP bounding boxes, run the command with the -b option:

python viewer.py -b <mcap_file>

To view the bounding boxes with an ONNX model, include the model with -m option:

python viewer.py -m <ONNX_model> <mcap_file>

Both sets of bounding boxes can be combined:

python viewer.py -m <ONNX_model> -b <mcap_file>

This project is licensed under the AGPL-3.0 or under the terms of the DeepView AI Middleware Commercial License.

Commercial Support is provided by Au-Zone Technologies through the DeepView Support site.