@inproceedings {pub2902,
	title = {Optical Flow Field Features for Audio-visual Word Prominence Detection},
	author = {Andrea Schnall AND Martin Ernst Heckmann},
	year = {2015},
	abstract = {In this paper we investigate visual features for the automatic labeling of the prominence of words. 
Visual motion is a rich source of information. Modifying the articulatory parameters to raise the prominence of a segment of an utterance, is usually accompanied by a stronger and different movement of mouth and head. One way to describe such motion is to use optical flow fields.
During the recording of the underlying audio-visual database, the subjects were asked to make corrections for a misunderstanding of a single word of the system by using prosodic cues only, which created a narrow and a broad focus.
Audio-visual recordings with a distant microphone and without visual markers were made. As acoustic features duration, intensity, fundamental frequency and spectral emphasis were calculated. From the visual channel the mouth region were extracted. From this region the optical flow were calculated and all the optical flow fields for one word were summed up. 
We show that using those features in addition to the audio features improve the classification results. 
The results are not as good as our former image transformation based visual features, but using both in addition to the audio features leads to the overall best results.},
	publisher = {IEEE},
	booktitle = {Int. Joint Conf. on Neural Networks (IJCNN)}
}