When a computer processes a visual image, it’s like a museumgoer standing too close to a pointillist painting: all it sees are rows of colored dots. But the next step forward for video processing could, in effect, be a step back; new digital technologies are beginning to discriminate objects.
Object discrimination presents the possibility of “saving a lot of bandwidth,” says Eran Yarom, CEO of Tel Aviv, Israel-based startup EnQuad Technologies. EnQuad has developed an algorithm that analyzes digital video streams and “extracts the object from the background in real time,” says Yarom. “Then we’re able to send the background and the moving object separately.” Video from the Winter Olympics, for instance, might feature a skater, an ice rink and an audience. The position of the skater needs to change with every frame of video; but transmit the rink and the spectators at a lower resolution, with less frequent updates, and you can cut your bandwidth requirements by three-quarters. EnQuad’s first product will be a digital signal processor that performs such selective transmission; Yarom expects it to reach the market by mid-2002.
Another Israeli startup is taking a hardware approach. A video camera from 3DV Systems of Yokneam, Israel, uses an infrared laser to gauge depth; existing graphics software can use that depth information to extract objects from the video stream. The startup’s vice president, Arend Verweij, says that the camera could be used, for example, for high-resolution videoconferencing.
Object extraction techniques can also help determine what’s going on in a moving image. John Clark, vice president of product management at ObjectVideo in Reston, VA, thinks the technology’s best short-term prospect is in video surveillance. “Surveillance cameras are everywhere,” Clark says, “recording a whole lot of valuable information, and in the vast majority of cases there’s nobody there to watch it.” ObjectVideo’s software could alert building security guards if a visitor entered a restricted area, for example-or left behind a bomb-sized briefcase on the way out. The company hopes to have a commercial product available by the end of the year.
Industry experts predict that these and similar techniques will ultimately carve up all video on the Web-at which point Britney Spears will really be the most objectified girl on earth.
The big new idea for making self-driving cars that can go anywhere
The mainstream approach to driverless cars is slow and difficult. These startups think going all-in on AI will get there faster.
Inside Charm Industrial’s big bet on corn stalks for carbon removal
The startup used plant matter and bio-oil to sequester thousands of tons of carbon. The question now is how reliable, scalable, and economical this approach will prove.
The dark secret behind those cute AI-generated animal images
Google Brain has revealed its own image-making AI, called Imagen. But don't expect to see anything that isn't wholesome.
The hype around DeepMind’s new AI model misses what’s actually cool about it
Some worry that the chatter about these tools is doing the whole field a disservice.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.