Emergency vehicle classification
Budget: $3,000 – $5,000 USD
Creation of ONNX classification model that loads in latest OpenCV and can effectively classify emergency vehicles from top front/top front side view (camera). Visual greyscale image of vehicle and audio processing of siren can be used.
Classification has to complete within 200ms on 500 GFLOPs.
Error rate:
0.97 positive for Emergency classification (1/30 emergency vehicles can be incorrectly classified as non-emergency).
0.99 positive for Non-emergency classification (1/100 of vehicles can be incorrectly classified as emergency).
You will have to create your own dataset but will also be provided RTSP access to a camera with frequent ambulance appearances.
This has to be done for ambulance and fire trucks marked according to specific standard (provided in attachment).
All data with source code needs to be handed over.
Happy to add any information that is required.
Classification has to complete within 200ms on 500 GFLOPs.
Error rate:
0.97 positive for Emergency classification (1/30 emergency vehicles can be incorrectly classified as non-emergency).
0.99 positive for Non-emergency classification (1/100 of vehicles can be incorrectly classified as emergency).
You will have to create your own dataset but will also be provided RTSP access to a camera with frequent ambulance appearances.
This has to be done for ambulance and fire trucks marked according to specific standard (provided in attachment).
All data with source code needs to be handed over.
Happy to add any information that is required.