TapTapSee, one of the apps that form part of the SilVer project, is one of the most appropriate for describing objects and scenes. The use it makes both of the computer image recognition system and of artificial intelligence has for some time provided the possibility to obtain exact descriptions of all that surrounds us.
In addition, TapTapSee offers the chance to process short videos rather than photos alone; this increases the precision with which a scene or object may be described. Sometimes, when a photo is taken and sent to the TapTapSee recognition servers to be identified, the angle of the snapshot means that the information we receive is not detailed enough. However, if instead of sending a photo, a video of the object or scene in question is recorded, the app will receive more data.
How TapTapSee uses video detection?
In the main interface of TapTapSee, in addition to the keys we already know, there is one that VoiceOver verbalises as “Video on”. When this is pressed, a series of whistles are heard. These signal the start of a video recording of a maximum of ten seconds.
While the video is being made, it is important for the camera to focus on whatever we wish to recognise. This should always be done from different angles in order to provide the app with a greater amount of information. When filming ends, TapTapSee processes the result and sends the entire description.