Demos

This video is an example of the real-time performances of the object detection capabilities achievable with a Tensor Flow Lite model (SSD MobileNet) in execution on the Google Edge TPU USB accelerator attached to a Raspberry Pi. Object labels and confidence scores have been overlaid in real-time while capturing the video (H264 1536x768).

A couple of demos which show a Raspberry Pi running a TensorFlow Light SSD model for 1) face and 2) glass + bottle real-time detection and tracking with the UV4L software. The little robot is in the middle of the table. Video is streamed from the robot to a browser on a laptop (on its left) in the same LAN through WebRTC (the display on the table is connected to the laptop, NOT to the robot). Everything runs on the Raspberry Pi side, thus no third-party services are required.

A simple demo which shows a Raspberry Pi 3B+ running a game (Super Mario) with RetroPie. Both the game audio and video are rendered in real-time from an html5 webapp running in Chrome on the Android smartphone. The gamepad which controls the player is connected to the smartphone via bluetooth. No plug-in is required by the browser. (For demo purposes a display is connected to the Raspberry Pi for, but this is not necessary, as the Raspberry Pi can be completely headless).

This is a demo which shows that it's possible to play RetroPie games in any browser (audio + video + keyboard control) with native standard web technologies because the imperceptible latencies involved.

Immersive telepresence! List of features: *. optionally two-way, encrypted live data-audio-video in HD or greater resolution at full frame rate between two participants (or even more) * FPV in a 180° x 180° panorama * 3D Stereo image (optional) * real-time head tracking, working out-of-the-box with any smartphone (no special software required, e.g. plugin-free browser) inside a VR headset like Google cardboards or easily adaptable to dedicate hardware like Oculus or OSVR. * over Internet (cloud) or LAN * does not necessarily require any third-party server infrastructure… * …but can be optionally easily made as on-demand service when put in the cloud * easy NAT traversal * audio-video on-the-fly recording & play back on the browser * client or recording on server side for further post-processing * adaptive bitrate streaming to available bandwidth

Raspberry Pi screen sharing from the PC browser (plugin-free!). This is an example of what it's possible to do with the UV4L Streaming Server and the raspidisp driver which turns the HDMI output in a low-latency Video4Linux2 capture device.

UV4L Demo to show *hardware-encoded* HD Video & Audio RealTime Streaming from RaspberryPi to browser with WebRTC.

Control your robot from the browser with a smartphone! As you can see the device orientation information is sent from the browser to the robot and echoed back to the browser in real time. Of course you can add (two-way) audio/video along with the data (not shown in this short demo).

Full HD Screen & Audio real-time mirroring from browser to RaspberryPi

Low latency streaming to Firefox of hardware-encoded video