Emodrive: Driver monitoring system that continuously monitors the emotion of the driver, speech in the vehicle and a few external conditions to quantify the safety of driving
Emotions which we consider are critical - Anger, Disgust and Fear.
- IBM Watson speech recognition API
- Algorithmia Image Analysis
The project was developed using Django Web framework in Python.
By feeding the image and Audio of the driver. We generate a safety score which is a measure of the combined emotions such as Anger, disgust and fear. It then presents a text message on the web page to display the next course of action for the driver. Messages can be, "Drive slow", "Call to 911" etc.
- Watson API's takes a lot of time to respond with the speech analysis.
- Incorporate sensors to get environmental factors such as climate, car speed etc.
- Incorporate video analysis over Image analysis
- Noise reduction to focus on Driver speech and video feed.
- Response module to account for next best action that can be taken by the driver.
- Work around for low Internet coverage problems.