Dear bosses, can help a friend to find a code for segmenting video semantic, because didn't contact before, so I do not know how to start, hope to draw lessons from with the bosses can share it