Our team has been working with Misty for a little while now, and are in need of video and audio data streams from Misty. I found this youtube video on the MISTY ROBOTICS channel, and it even shows a live video stream from Misty. I know this isn’t a new request as others have posted about it. We’re looking to do some real-time (or as close as we can get to it) image recognition and speech transcription (via Google Cloud) for some of our current research.
We’ve looked into temporary workarounds such as setting the camera on a separate thread and polling at set intervals to simulate a low framerate video. This quickly proved to be a poor solution, as we couldn’t go sub-one second intervals (we’re using the REST API, and the Python Wrapper). We also couldn’t figure out a way to accurately start/stop the audio recording when a person started/stopped talking to Misty aside from using the wake word, which won’t be acceptable for our research. For now, we’ve circumvented these issues using external cameras and microphones, but it would be awesome if we could use the ones built in to Misty!
I was curious if the community has some proposed/tested solutions to this problem, and would love to hear about them! If you are also in need of access to the data streams, there is a post on the wishlist regarding this topic. So be sure to show the developers you desire this feature.