Originally published on Jul 27, 2016

With tagged photos you can see who is who and where, but for a video the current way of tagging doesn’t work very well. So I thought of a concept about how to use Image Recognition to tag people in videos.

When you select to tag a video, it could be automatically scanned to recognize and show persons like you and your friends appearing in video. Then you can choose which ones to tag. If there’s someone which couldn’t be recognized you could add it manually and let this face be recognized throughout video.

Anyone watching the video could then select a tag to watch only the parts where this person or any tagged image appears.

Besides the face, there could be an option to tag the voice, too. It could be associated automatically which voice is belonging to which person. If not, you could tag it manually, as well. So when you select to watch someone, you wouldn’t miss any part where the person speaks but doesn’t appear in video.

Other than tagging you could also search someone other than friends in videos and fast-forward to his/her parts.

People mostly use Facebook to post their videos, so I've chosen it as context and adapted the idea to its interface.

Things have changed since when I made this in 2016, so Facebook is not the right context anymore. But I think the idea of being able to search/browse through videos -not only people but for any recognized object- is still relevant and applicable, especially with the ever increasing use of video content and AI.

Originally published on Jul 27, 2016

With tagged photos you can see who is who and where, but for a video the current way of tagging doesn’t work very well. So I thought of a concept about how to use Image Recognition to tag people in videos.

When you select to tag a video, it could be automatically scanned to recognize and show persons like you and your friends appearing in video. Then you can choose which ones to tag. If there’s someone which couldn’t be recognized you could add it manually and let this face be recognized throughout video.

Anyone watching the video could then select a tag to watch only the parts where this person or any tagged image appears.

Besides the face, there could be an option to tag the voice, too. It could be associated automatically which voice is belonging to which person. If not, you could tag it manually, as well. So when you select to watch someone, you wouldn’t miss any part where the person speaks but doesn’t appear in video.

Other than tagging you could also search someone other than friends in videos and fast-forward to his/her parts.

People mostly use Facebook to post their videos, so I've chosen it as context and adapted the idea to its interface.

Things have changed since when I made this in 2016, so Facebook is not the right context anymore. But I think the idea of being able to search/browse through videos -not only people but for any recognized object- is still relevant and applicable, especially with the ever increasing use of video content and AI.