Video sequence ID by decimated scene signature

An apparatus for identifying video content, the apparatus comprising:

• a scene detector to detect a plurality of repeatably detectable scenes in the video content;
• a signature generator to determine a scene signature for each of the plurality of scenes, each scene signature including concatenated pixel data from a repeatable selection of pixels in a repeatably selected frame in the video content, where the selection of pixels represent a subset of less than all pixels in the frame, and where the pixel data includes a truncated segment of pixel data;
• and
• a content identifier
• to provide a plurality of scene signatures to a comparator that compares the plurality of scene signatures to a database having a plurality of video content items, the database including previously determined scene signatures for each item of video content in the database, and
• to receive an identification of the video content from the comparator;
• wherein the repeatably selected frames are predetermined frames following scene changes, and each of the frames has the same temporal position after the occurrence of its related scene change.

A method for identifying video content, the method comprising:

• receiving a plurality of scene signatures, each scene signature including concatenated pixel data from a repeatable selection of pixels in a repeatably selected frame in the video content in the video content, where each selection of pixels represents a subset of less than all pixels in the frame from which the pixels are selected, and where the pixel data includes a truncated segment of pixel data;
• determining an identification of the video content by matching the plurality of scene signatures to a database having a plurality of video content items, the database including previously determined scene signatures for each repeatable sequence of frames for each item of video content in the database, and
• providing the identification of the video content;
• wherein the repeatably selected frames are predetermined frames following scene changes, and each of the frames has the same temporal position after the occurrence of its related scene change.

A method for identifying video content, the method comprising:

• identifying a scene in the video content;
• identifying a repeatably selected frame in the scene;
• selecting a plurality pixels in the selected frame using a pixel selection algorithm, the pixels being selected to provide a repeatable selection of pixels in the video content, where the repeatable selection of pixels represents a subset of less than all pixels in the selected frame, and where the pixel data either includes truncated pixel data or eliminates certain pixel data;
• establishing a scene signature by:
• extracting a characteristic of each selected pixel that can be represented in a digital form as pixel data for each pixel, truncating the digital form for each pixel, and generating a digital stream by concatenating the digital forms for the extracted characteristic of each pixel;
• transmitting the scene signature to an identification device for comparison with scene signatures in a database having a plurality of video content items to identify the video content;
• and
• receiving either the identification of the video content or an indication that no video content is uniquely identified by the scene signature;
• wherein the repeatably selected frames are predetermined frames following scene changes, and each of the frames has the same temporal position after the occurrence of its related scene change.