Twitch announced that it is to implement refined search functionality, especially for Overwatch and Hearthstone fans. According to the released information, players will be able to “filter channels by hero played on Overwatch streams, or by game mode, player rank, number of wins, and hero class on Hearthstone streams.”
The new functionality is considered an ultimately convenient tool for broadcasters who would like to target a particular audience and demonstrate how they perform in a certain game. Previously, a metadata of this kind was introduced for League of Legends and it has achieved a great success.
JT Gleason, Director of Integration Success at Twitch, said that Overwatch and Hearthstone are two of the most popular games and players would be eager to polish their skills by watching others play. Consequently, the development team has decided to focus on improving the discovery and enable fans to find “more granular aspects of the gameplay”.
The refined search feature is a courtesy of ClipMine. The latter is a learning-based video indexing platform Twitch has recently acquired. Essentially, ClipMine translates visual information into frame-level metadata. Then, the metadata is utilized for video analyzing and improving discovery across the platform.
Zia Syed, the ClipMine founder, commented on the new Overwatch and Hearthstone filter addition. He was quoted as saying that the specifics of the video game streams’ structure makes them difficult to exploit as far as content discovery was concerned. Nevertheless, ClipMine managed to recreate the structure in a “reliable (…), cost effective manner” thus enabling Twitch users to explore the content in a new way.
The new feature is expected to be fully implemented in the next few hours. For the time being, it remains unknown whether other games will get similar functionality. Twitch viewers remain hopeful that CS:GO will also receive a feature extension sometime in future.