February 9, 2015
Startup Clarifai has developed artificial intelligence technology based on deep learning that can identify what is in a video. This ability could be significant for search engines, which currently have to rely on textual clues around a video to guess what might be in it. Clarifai’s AI has the ability to identify objects, in addition to letting users know exactly when those objects will appear in the video. This technology could be used to help advertisers and other companies analyze their videos.
Clarifai’s AI can identify objects in videos based on deep learning. The structure of the software is built like the neural networks in a human brain. The AI is trained to derive information from certain images and audio, and then make inferences about content that contains similar clips of media. The more videos the AI analyzes, the better it will get at identifying objects in them.
Matthew Zeiler, Clarifai’s founder and chief executive, believes the technology is only the tip of the iceberg. He’s already working with advertisers to help them analyze video and figure out when is a good time to drop an ad into the video, according to VentureBeat.
One day, it may also make YouTube searches more effective, allow sports videographers to quickly locate the best footage, and maybe even send alerts by analyzing security footage. The process only takes a 10 seconds to analyze three minutes of video.
Clarifai already has some experience in the image recognition business. “Clarifai already is selling high-quality image search software that could be used on smartphones, by catalog companies, or anyone who needs a large number of images,” Wired reports.
Zeiler’s AI algorithms took home the top prize at the prestigious ImageNet competition.