Who would have imagined 10 years ago that today we would have computer programs that can differentiate moving cats from dogs in a video in matter of milli-seconds!?
Joseph Redmon works on a system called YOLO (You Only Look Once), which is an open-source method of object-detection that can super-quickly detect many sorts of objects in images and videos.
The implications of this technology can be in may fields such as self-driving cars, robotics, cancer detection and so on.
Checkout the live demo of this application in this TED talk:
How does YOLO work?
The YOLO algorithm applies a single neural network to the full image. This network divides the image into regions and predicts bounding boxes and probabilities for each region. These bounding boxes are weighted by the predicted probabilities. It is +1000x faster than R-CNN and 100x faster than Fast R-CNN. See this paper or the YOLO website for more details on the full system.