All Downloads are FREE. Search and download functionalities are using the official Maven repository.

org.openimaj.video.processing.timefinder.ObjectTimeFinder Maven / Gradle / Ivy

/**
 * Copyright (c) 2011, The University of Southampton and the individual contributors.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modification,
 * are permitted provided that the following conditions are met:
 *
 *   * 	Redistributions of source code must retain the above copyright notice,
 * 	this list of conditions and the following disclaimer.
 *
 *   *	Redistributions in binary form must reproduce the above copyright notice,
 * 	this list of conditions and the following disclaimer in the documentation
 * 	and/or other materials provided with the distribution.
 *
 *   *	Neither the name of the University of Southampton nor the names of its
 * 	contributors may be used to endorse or promote products derived from this
 * 	software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
/**
 * 
 */
package org.openimaj.video.processing.timefinder;

import java.util.List;

import org.openimaj.image.Image;
import org.openimaj.math.geometry.shape.Rectangle;
import org.openimaj.util.pair.IndependentPair;
import org.openimaj.video.Video;
import org.openimaj.video.VideoCache;
import org.openimaj.video.processing.tracking.ObjectTracker;
import org.openimaj.video.timecode.HrsMinSecFrameTimecode;
import org.openimaj.video.timecode.VideoTimecode;

/**
 * 	This class is a general class for finding the time range in which a specific
 * 	object is found within a video.  Given a seed frame in which the object appears,
 * 	this class will track forwards and backwards in the video stream to find when
 * 	the object disappears. It is then able to provide the timecodes at which the
 * 	object can be found within the video.
 *
 *	@author David Dupplaw ([email protected])
 *	
 *	@created 14 Oct 2011
 */
public class ObjectTimeFinder
{
	/**
	 * 	An interface for objects that wish to be informed of object tracking
	 * 	events during the time finder's execution.
	 * 
	 *  @author David Dupplaw ([email protected])
	 *	
	 *  @param  Type of object tracked
	 *  @param  Type of {@link Image} 
	 *	@created 7 Nov 2011
	 */
	public interface TimeFinderListener
	{
		/**
		 * 	Called when an object is tracked in the video.
		 * 
		 *  @param objects The objects being tracked
		 *  @param time The timecode of the frame in which the object was found.
		 *  @param frame The frame in which the object was found.
		 */
		public void objectTracked( List objects, VideoTimecode time, I frame );
	}
	
	/**
	 * 	Given a video, a keyframe (timecode) and a region of the image, 
	 * 	this method will attempt to track the the contents of the rectangle
	 * 	from the given frame, back and forth to find the place
	 * 	at which the object appears and disappears from the video.
	 *
	 *	@param  The type of image returned from the video
	 *	@param  The type of object being tracked
	 *	@param objectTracker The object tracker that will track the object for us
	 *	@param video The video 
	 *	@param keyframeTime The keyframe timecode to start at
	 *	@param bounds The bounding box of the object you wish to track
	 *	@param listener A listener for object tracking events
	 *	@return An {@link IndependentPair} of timecodes delineating the start
	 *		and end of the video which contains the object
	 */
	public ,O> 
		IndependentPair trackObject(
				ObjectTracker objectTracker, Video video, 
				VideoTimecode keyframeTime, Rectangle bounds, 
				TimeFinderListener listener )
	{
		// Set up the initial start and end timecodes. We'll update these
		// as we scan through the video
		HrsMinSecFrameTimecode startTime = new HrsMinSecFrameTimecode( 
				keyframeTime.getFrameNumber(), video.getFPS() );
		HrsMinSecFrameTimecode endTime = new HrsMinSecFrameTimecode( 
				keyframeTime.getFrameNumber(), video.getFPS() );
		
		// Now set the video to the start frame
		video.setCurrentFrameIndex( startTime.getFrameNumber() );
		I image = video.getCurrentFrame();
		I keyframeImage = image.clone();
		
		// Initialise the tracking with the start frame
		objectTracker.initialiseTracking( bounds, image );
		
		// Now we'll scan forwards in the video to find out when the object
		// disappears from view
		boolean foundObject = true;
		while( foundObject && image != null )
		{
			// Track the object
			List objects = objectTracker.trackObject( image );
			
			// If we've found no objects in this frame, then we've lost
			// sight of the object and must stop
			if( objects.size() == 0 )
				foundObject = false;
			else
			{
				// Move on to the next frame
				image = video.getNextFrame();
				
				// Update the end time
				endTime.setFrameNumber( video.getCurrentFrameIndex() );
				
				if( listener != null )
					listener.objectTracked( objects, endTime, image );
			}
		}
		
		// Reinitialise the object tracker to our start point
		objectTracker.initialiseTracking( bounds, keyframeImage );
		foundObject = true;

		// Now we're going to scan backwards through the video. However, scanning
		// backwards can be really slow, so what we're going to do is scan back
		// by a big chunk and then cache that chunk, before we track back through
		// it frame by frame. This reduces the number of times we need to do the
		// expensive backwards track.
		int nFramesToJumpBack = (int)video.getFPS()*2;	// We'll go 2 seconds at a time

		// currentTimecode will give the start of the chunk
		HrsMinSecFrameTimecode currentTimecode = new HrsMinSecFrameTimecode( 
				Math.max( startTime.getFrameNumber() - nFramesToJumpBack, 0 ), 
				video.getFPS() );

		// Now keep looping until we either lose the object or 
		// we hit the start of the video
		while( foundObject && currentTimecode.getFrameNumber() >= 0 )
		{
			// Slightly confusingly here, the startTime is the end of the chunk
			VideoCache vc = VideoCache.cacheVideo( video, currentTimecode, startTime );
			
			// Now track back through the cached video
			for( int n = 1; n <= vc.getNumberOfFrames(); n++ )
			{
				// Track backwards through the frames
				image = vc.getFrame( vc.getNumberOfFrames()-n );
				
				// Track the object
				List objects = objectTracker.trackObject( image );
				
				// If we've found no objects in this frame, then we've lost
				// sight of the object and must stop
				if( objects.size() == 0 )
				{
					foundObject = false;
					break;
				}
				else
				{
					// Update the start time
					startTime.setFrameNumber( startTime.getFrameNumber()-1 );
					
					if( listener != null )
						listener.objectTracked( objects, startTime, image );
				}			
			}

			// If we're already at frame zero, we should end this loop
			if( currentTimecode.getFrameNumber() == 0 )
				foundObject = false;
			
			// Update the chunk timecode if we need to keep going
			currentTimecode.setFrameNumber( 
					Math.max( startTime.getFrameNumber() - nFramesToJumpBack, 0 ) );
		}
	
		// Return the videos
		return new IndependentPair( startTime, endTime );
	}
}