HIGH 5! Here is the news update of the latest trends from the global technology and startup scene. HIGH 5 is published by Lab1886. This week’s topics are:
AI manages street signs using Google Street View +++ MIT robots acquire a sense of touch +++ More sustainable batteries +++ Cities with big carbon footprints +++ AI “awakens” surveillance cameras
For municipalities, locating damaged street signs is usually an awkward, costly process. Now researchers at Australia’s RMIT University have developed a fully automated system in which AI-based object detection analyzes street signs in freely available images such as Google Street View. Using 2D images, the system can detect signs with 96 percent accuracy and geolocate them with 98 percent accuracy. The system should be easy to scale for municipalities and transportation agencies.
Scientists at MIT’s Computer Science and AI Lab (CSAIL) have developed an AI that can “feel” an object’s tactility just by seeing it. If the AI is connected to a Kuka robot arm, it can also generate a visual representation of an object just by touching it. In exhaustive training sessions, the AI system handled 200 different household items.
Source: The Next Web
Driving vehicles powered by renewable energy is not the only challenge: EV production should also become more sustainable. Raw material extraction should be greener – in Australia and Canada, mining for cobalt and lithium is more strictly regulated than ever. In any case, use of cobalt – currently a key mineral – could be reduced or even completely eliminated. Production processes should also become more energy-efficient.
Source: heise.de (German)
Although only half of the world’s population lives in conurbations, just 100 cities are responsible for 18 percent of global carbon emissions. Big cities with the worst carbon footprints consume as much as small countries. This is according to a study conducted by researchers from various universities who examined 13,000 cities. They have presented their findings on an interactive map which can be used to look up cities worldwide.
Data is collected from millions of video feeds, but manual analysis often lags (far) behind. A new generation of video analytics software can analyze the movements and behavior of pedestrians in real time and report “anomalies” such as arguments or drunkenness. It can also detect, for example, vehicles driving in the wrong direction.
HIGH 5 is published by Lab1886. If you have any questions or suggestions, please contact us using the following email address: David Menzel (email@example.com)