DataFission has developed generally applicable technology that can index, search and report on all forms of unstructured, metadata-free digital data. With deep roots in military research, this San Jose company has developed algorithms that can index files and streams of data, without human interaction, in one sweep. The system can index any type of data, including video, still images, music, audio tracks, text, software and metadata. According to its brochure, DataFission is “Solving the world’s problem of searching the unsearchable.”
Once indexed, the data can be searched and retrieved using Boolean logic. For example, a live football game video stream could be indexed in real time, and moments later the producer could pull the exact clip containing player X + football + end zone.
Music can be searched by matching a specific riff, and other instances of that riff and related riffs can be retrieved.
Malware can be detected in two ways. If a scan of a known file doesn’t match previous instances of the file, or if a file contains previously known malware, it can be flagged.
Similarly, DataFission can scan the Web, locate instances of unauthorized content, and supply links back to those instances.
Other possible uses of DataFission technology include search and monetization of media elements that don’t have associated metadata, content management, data reduction and summarization, and trend and predictive analysis.
DataFission is demonstrating its tech this week during NAB at Booth C2453B in the Las Vegas Convention Center.