Fujitsu has developed a compression algorithm that can reduce video sizes by 90%

Video accounts for the overwhelming share of cloud storage, consumes the most computing power and digital traffic in the world, so anything that helps curb its gigantic appetites will save a significant amount of time and money.

According to Fujitsu, its specialists have developed a new compression algorithm that can reduce video size by 90%. True, it should be clarified that the video obtained after such compression can only be recognized by artificial intelligence. The reason is the high degree of image degradation.

A key aspect of the new Fujitsu technology is the ability to automatically analyze and recognize images using AI. In practice, compression will be carried out at the periphery (that is, on the video device itself) using a specialized chip, and recognition will take place in the cloud. In this case, both processes will be combined into a closed loop.

The technology will enable users to analyze video data by combining information from multiple video sources, data from cameras or sensors, and various business data such as sales statistics in the cloud.

The need for this technology arose with the advent of ultra-high-definition cameras on smartphones (for example, the Samsung S 20 Ultra with its 108-pixel sensor) and 4K video surveillance cameras. The technology will become available to users at the end of 2020.