Method and device for describing and capturing video object

A method for capturing video objects, comprising:

• replaying video objects;
• obtaining a video object region identification number (RID) and a video object region tracking number (TID) corresponding to a tracked video object according to information carried in an obtained video object tracking sequence (OTS);
• searching a personal identification number (PID) and features of the tracked video object according to a video object descriptor (OD);
• tracking the video object RID corresponding to the video object region TID in subsequent frames;
• and
• obtaining consecutive video object region maps (ORMs) by matching the video object RID with video sequences, and marking the video ORMs in real time.

A device for describing video objects, comprising:

• a video object capturing module, adapted to:
• capture video images, generate video sequences, and export the video sequences to a video object tracking module and a video object analyzing module respectively;
• the video object tracking module, adapted to:
• generate a video object tracking sequence (OTS) according to received video sequences, and export the video OTS to the video object analyzing module;
• and
• the video object analyzing module, adapted to:
• generate video object descriptors (ODs) according to the video OTS exported by the video object tracking module and the video sequences exported by the video object capturing module.

A device for capturing video objects, comprising:

• a replay control module, adapted to:
• replay video objects, receive video objects selected by a user, match the video objects with video object region maps (ORMs) contained in a video object tracking sequence (OTS), and obtain a tracked video object region identification number (RID);
• query a video object region index (ORI) according to the tracked video object RID to obtain a tracked video object region tracking number (TID), track the video object RID related to the video object region TID in subsequent frames, and export the video object RID to a video overlay module;
• the video overlay module, adapted to:
• match the video object RID contained in received information with video sequences to obtain consecutive video ORMs, mark the video ORMs in real time, and export the video ORMs to a displaying module;
• and
• the displaying module, adapted to:
• display the marked video ORMs in real time.